Knowledge Centre · Water quality and aquarium stability

Aquarium maintenance schedule: what to do, how often, and when your tank should change the plan

There's a genuinely useful default rhythm for aquarium maintenance: test weekly, change part of the water on a regular cycle, clean filter media only when it needs it and never all at once, and keep a daily eye on your fish without disturbing anything. But the right cadence for your specific tank depends on stocking, feeding, plant growth and what your last few readings actually showed — not a date circled on a calendar. This guide sets out a sensible starting routine, what should make you deviate from it, and the over-maintenance mistakes that cause more instability than the problem they were meant to prevent.

Before you build a routine

A schedule works when…

You treat the frequencies below as a starting point, adjust them once you know your own tank's stocking, feeding and readings, and let a test result — not the calendar — decide whether something needs attention today.

A schedule stops working when…

You follow it rigidly regardless of what your readings show, do several thorough cleaning tasks on the same day "to get it all done," or treat a clean-looking tank as proof nothing needs testing.

Worth knowing: the single most common way a maintenance routine backfires isn't neglect — it's doing too much, too thoroughly, on the same day. See "Why doing everything at once can destabilise a tank" below.

A sensible default rhythm

Treat this as a starting point for a newly established, moderately stocked freshwater community tank — not a fixed rule. Every row below has a reason it might need to happen more often, covered in the sections that follow.

A default rhythm for an established, moderately stocked tank. Adjust using the sections below, not by following this table blindly.
How often	Task	Why this frequency
Daily	Look at your fish and check equipment is running	Behaviour and appetite change faster than most water-chemistry problems build.
Weekly	Test ammonia, nitrite, nitrate and pH; partial water change	Frequent enough to catch a drift before it becomes a crisis, without over-disturbing the tank.
Weekly or fortnightly	Clean the glass; spot-clean visible algae	Mostly cosmetic and safe to do often, though it won't fix a lighting or nutrient imbalance on its own — see "Glass, substrate and algae" below.
Monthly	Rinse mechanical filter media (sponge/floss) in removed tank water; check flow	Mechanical media clogs with debris faster than biological media needs disturbing.
Every one to three months	Check GH/KH; test for detectable phosphate if algae has been a recurring issue	These generally drift more slowly than ammonia, nitrite or nitrate in an established tank.
Rarely, and never all at once	Rinse or replace biological filter media; deep-clean substrate	Both disturb the surfaces your nitrifying bacteria actually live on — see "Filter maintenance" below.

Water testing

Weekly is the right default for an established, stable tank, but it's a floor, not a ceiling. Test more often — every day or two — while a tank is still cycling, for the first couple of weeks after adding new livestock, after any equipment failure or power cut, or whenever a fish looks or behaves differently. A single test only proves what a reading was at that moment; it's testing regularly and comparing results over time that actually shows whether your tank is stable, drifting, or recovering from something.

Which parameters matter most depends on what you're checking for: our ammonia, nitrite and nitrate guides cover the core nitrogen-cycle readings most tanks should test weekly, while pH and GH/KH generally drift more slowly once a tank is established and can usually be checked less often, unless you're actively adjusting either. If you haven't settled on a test kit yet, see our choosing a water test kit guide for liquid, strip and digital options compared.

Water changes

A weekly or fortnightly partial change suits many established, moderately stocked freshwater tanks well, but the right interval and volume depend on your stocking, feeding and how quickly nitrate builds up between changes — our aquarium water changes guide covers this in full, including why a heavily stocked, heavily fed or newly cycling tank often needs more frequent attention than a lightly stocked, well-planted one. Use the Water Change Assistant to work out roughly what a given change will do to a specific reading, rather than guessing at a percentage.

A consistent routine that keeps your baseline readings stable, or gently falling between changes, is more useful than chasing an exact percentage or schedule copied from somewhere else.

Filter maintenance

Split filter maintenance into two genuinely different tasks, because treating them the same is where most over-maintenance damage happens. Mechanical media — sponges, floss, anything trapping physical debris — clogs with waste and can be rinsed reasonably often, roughly monthly for many tanks, since its job is purely physical filtration. Biological media, and the surfaces of sponges and cartridges that have matured over time, carry the bacterial colony that actually processes ammonia and nitrite; disturbing that far more often, or far more thoroughly, than necessary is one of the most common ways an established tank ends up with a detectable ammonia or nitrite reading it didn't have the week before.

Whenever you do need to clean anything carrying biological growth, rinse it in water removed from the tank, never tap water — chlorine and chloramine are added specifically to kill microorganisms, and that includes the ones you're trying to keep. Check your filter's actual flow against what your tank and livestock need with the filter turnover calculator at the same time, since reduced flow from a partially clogged intake or impeller is itself a common, easily missed maintenance gap.

Glass, substrate and algae

Routine glass cleaning is safe to do often and rarely causes problems on its own, but it's worth understanding what it actually fixes: most algae is a lighting or nutrient-management issue, not a hygiene issue in the everyday sense, so scrubbing the glass without addressing light duration, intensity or excess nutrients tends to bring the same growth straight back. Our algae in aquariums guide covers identifying what you're looking at and the underlying causes worth correcting alongside routine cleaning.

Substrate needs a lighter touch than glass. Light gravel vacuuming during a routine water change is generally enough; a full, deep substrate disturbance is closer to the "rarely, and never alongside other major cleaning" category, for the same bacterial-surface-area reason covered below. It's also why an old, undisturbed patch can carry a genuine safety consideration if you dig into it all at once — see our guide to what an aquarium smell means if a deep clean has ever released an unexpected odour.

Equipment checks

A quick daily glance at your heater, filter and any powerhead confirms they're actually running, not just installed. Check a heater's actual output against a separate thermometer periodically rather than trusting its dial alone, since a stuck or failed heater is a genuine, sometimes fatal risk that isn't always obvious from a glance. Confirm lids and any equipment entry points are still genuinely gap-free, particularly for jumping-prone species — several species guides on this site, including current-loving species that need a secure lid, call this out specifically.

Livestock observation

A daily look at feeding response, colour, fin condition and general behaviour catches problems that testing alone won't, since a fish can be behaving differently well before a water-quality reading confirms a cause. If something looks off, our Fish Symptoms hub and Emergency Diagnosis tool can help you prioritise what to check first, rather than reacting to the first thing that comes to mind.

Planted tanks

Healthy, actively growing plants take up some ammonia and nitrate directly, which can ease the nitrate side of a maintenance routine compared with an unplanted tank, but they add tasks of their own: trimming back overgrowth, removing decaying leaves before they add to the waste load, and dosing or substrate-feeding anything with real nutrient demands. A planted tank doesn't need testing any less often than an unplanted one — it can still drift or crash, sometimes just more slowly, which can make a genuine problem easier to miss for longer rather than less likely to happen.

Why doing everything at once can destabilise a tank

Most of a tank's nitrifying bacteria live as a biofilm on filter media, substrate and other submerged surfaces, not free-floating in the water[1]. Replacing all of a filter's biological media at the same time, rinsing it in chlorinated or chloraminated tap water, or deep-cleaning substrate and scrubbing every decoration on the same day can remove a large share of that population in one go — and it's a well-documented cause of what hobbyists call a cycle crash: a detectable ammonia or nitrite reading appearing in a tank that had previously tested clean for weeks or months[1].

The instinct behind a really thorough clean is understandable, but it's usually the wrong one. Leaving more of the tank's surface area undisturbed at any one time — spreading mechanical cleaning, biological media attention and substrate work across different weeks rather than tackling all of it in one session — is the safer default, even though it feels like less is getting done.

When to deviate from any schedule

Every frequency on this page is a starting point for an established, moderately stocked tank. Test and change water more often than the default while a tank is still cycling, for the first few weeks after new livestock arrives, after treating with medication (some treatments affect filtration or require carbon removal — check the product's own instructions), or whenever a reading has recently been outside its normal range for your tank. Conversely, a long-established, lightly stocked, heavily planted tank with months of consistently stable readings may reasonably need less frequent testing than a newly set-up or heavily stocked one — but let that track record, not an assumption, be what tells you it's safe to ease off.

Common mistakes

Cleaning the filter, substrate and glass thoroughly on the same day — each is individually reasonable; combined, they can remove more of the bacterial population than any one task would alone.
Rinsing filter media in tap water, exposing it to chlorine or chloramine added specifically to kill microorganisms.
Treating a clean-looking tank as proof nothing needs testing — ammonia, nitrite and a slow nitrate creep aren't reliably visible.
Doing an oversized water change in response to a minor reading, which can shock livestock with a large, sudden shift in water chemistry rather than correcting the issue gradually.
Following a fixed calendar regardless of what recent readings actually show, in either direction — too rigid to catch an unusual week, or too relaxed once a routine feels established.

Frequently asked questions

How often should I really test my water?

Weekly is a sensible default for an established, stable tank. Test more often — every day or two — while a tank is cycling, shortly after adding new livestock, after any equipment problem, or if a fish looks or behaves unusually. Once you have months of stable weekly readings behind you, some keepers stretch the interval slightly, but the reading itself should be what tells you it's safe to do that, not the calendar.

Is it bad to clean my filter and do a big water change on the same day?

It can be, particularly if "cleaning the filter" means rinsing or replacing all of its biological media at once. Most of your tank's nitrifying bacteria live on filter media, substrate and other submerged surfaces rather than free-floating in the water, and disturbing several of those surfaces heavily on the same day removes more of that population than any one task would alone. Spread thorough cleaning tasks across different days where you can, and never rinse filter media in tap water.

Can you over-maintain an aquarium?

Yes. Over-cleaning is a well-documented way tanks end up unstable — scrubbing every surface, replacing all filter media at once, or doing an oversized water change in response to a minor reading can remove more of the beneficial bacterial population than the original issue justified, or shock livestock with a large, sudden change in water chemistry. Thorough is not the same as safe; disturbing less of the tank at once is usually the better instinct.

My tank looks completely fine — do I still need a maintenance routine?

Yes, and this is precisely when routine testing earns its keep. Ammonia, nitrite and a slow nitrate creep are not reliably visible — a tank can look completely normal while a reading has already drifted, and by the time something looks wrong, the underlying issue may have been building for a while. Looking fine is a reasonable sign nothing urgent is happening; it isn't a substitute for actually testing.

Do heavily planted tanks need a different maintenance routine?

In some ways, yes. Healthy, fast-growing plants take up some ammonia and nitrate directly, which can ease the nitrate side of the picture compared with an unplanted tank, but they add their own tasks — trimming, removing decaying leaves, and dosing or substrate fertilisation if you're growing anything demanding. Testing frequency doesn't drop just because a tank is planted; a planted tank can still crash or drift, just sometimes more slowly.

Holding the pattern somewhere other than memory

A weekly test is easy enough to remember this week. The harder part is holding the whole pattern in your head six months from now — whether this Tuesday's nitrate is actually higher than it's been running, whether it's really been four weeks since the filter media was last touched, whether last month's water change actually happened or just felt like it did. Logging your readings and routine in AquaPulse keeps the pattern somewhere other than memory, so a slow drift or a slipped task is something you can check against, not something you're relying on having noticed.

AquaPulse works from the readings and context you enter — it doesn't run the tests for you or track maintenance automatically unless you tell it. For the plain-language explainer without the app, see the Water Check feature page.

Sources and editorial information

Published: 7 August 2026
Last reviewed: 7 August 2026
Written and maintained by: AquaPulse (Valeon Labs Ltd)

This guide is written and maintained by the AquaPulse team at Valeon Labs Ltd. It synthesises guidance already researched and sourced across our own water-quality, cycling and algae guides — see the "Related guidance" section below and the individual guides linked throughout — rather than introducing new factual claims. Where a frequency is described as a "default" or "starting point," that reflects widely repeated hobby practice, not a fixed rule every tank must follow; we've been explicit about that throughout rather than presenting one schedule as universally correct. We'll update this page if we find something we've got wrong.

Good to know: This guide provides general informational guidance only. It is not a substitute for testing your own water with a reliable kit, or for specialist aquatic-health advice for serious or persistent livestock health issues.

[1] University of Florida IFAS Extension — "Introduction to Fish Health Management" (CIR921/FA004) General fish-health-management extension source, consistent with this publication's use elsewhere on AquaPulse. Used here for the general principle that nitrifying bacteria are surface-associated (biofilm) rather than free-floating, and that disturbing those surfaces can reduce biological filtration capacity — the same underlying mechanism our own cycle crash guide covers in full, which this page links to rather than re-deriving.

What we deliberately left out: a single universal maintenance calendar. Stocking density, feeding, plant load, tank age and livestock all genuinely change the right frequency, and presenting one fixed schedule as correct for every tank would contradict the guidance in this page's own body text.