The rule of thumb, stated plainly
The figure in general circulation is about 20 BTU per hour per square foot of floor area, for a room with an eight foot ceiling in average conditions. Three hundred square feet gives 6,000 BTU/h, which is why 5,000 and 6,000 BTU window units exist in the volume they do. The common adjustments layered on top of it are equally standard: scale for ceiling height, add or subtract about ten percent for heavy sun or heavy shade, add roughly 600 BTU/h for each occupant beyond two, and add about 4,000 BTU/h if the room is a kitchen.
All of those numbers are conventions, not measurements. Different sources quote 18, 20 or 25 BTU per square foot, and the per-person and kitchen adders vary too. That is why the rate is an editable field here rather than a constant: if you have a figure you trust for your climate or your building type, use it. The output is a range because the input is a convention, and a range communicates the uncertainty honestly where a single number to the nearest hundred BTU would not.
Why too big is worse than slightly too small
An air conditioner does two jobs. It lowers the air temperature, and it removes moisture from the air by condensing it on a cold coil. The first job happens quickly. The second only happens while the compressor has been running long enough for the coil to get properly cold and for a sustained volume of air to pass across it.
An oversized unit does the first job so fast that it never gets to the second. It blasts the room down to setpoint in a few minutes, the thermostat is satisfied, it shuts off, the room warms back up, and it starts again. The temperature reading is correct and the room feels wrong — cool but clammy, because the humidity was never removed. That sensation drives people to set the thermostat even lower, which makes everything worse.
| Slightly undersized | Correctly sized | Oversized | |
|---|---|---|---|
| Run pattern | Runs almost continuously in peak heat | Long cycles, mostly satisfied | Short bursts, frequent restarts |
| Humidity | Well controlled | Well controlled | Poorly controlled — cold and damp |
| Peak performance | May hold a degree or two above setpoint on the worst day | Holds setpoint | Holds setpoint |
| Equipment life | Fine | Fine | Compressor cycling wears it faster |
| Comfort | Good, steady | Good, steady | Temperature swings, noise on every start |
The consequence of undersizing shows up on perhaps a handful of days a year and is a mild disappointment. The consequence of oversizing shows up every single day the unit runs. When the estimate falls between two available sizes, that asymmetry is the argument for the smaller one.
What the square-footage rule cannot see
Cooling load is heat arriving in a room, and floor area is a weak proxy for it. Consider what actually drives the number. Solar gain through glazing is often the largest single term, and it depends on how much glass there is, which direction it faces, whether it is shaded, and what the glass is — a west-facing wall of single glazing and a small north window in the same square footage are not comparable. Conduction through the envelope depends on insulation levels and on what is on the other side: a room under an uninsulated roof in the sun picks up an enormous amount that a room with conditioned space above does not.
Then there is infiltration, which is outdoor air leaking in and having to be cooled and dried, and which varies by more than a factor of five between a tight modern build and a drafty old house. Then internal gains: people, lighting, a computer, a television, a gaming machine, and in a kitchen the cooking equipment. And underneath all of it, the local design temperature — the same room needs a different machine in Phoenix and in Portland.
None of that is visible in a length times a width, which is why the answer above is a place to start shopping rather than a specification.
What a load calculation actually is
When an HVAC contractor sizes equipment properly, they perform a room-by-room heat gain calculation using a standard method — in the US, ACCA Manual J is the recognised one, and building codes in many jurisdictions require it for permitted installations. It takes the actual construction of each wall, window, floor and ceiling, the orientation, the shading, the measured or estimated infiltration rate, the internal gains and the local design conditions, and produces sensible and latent load figures in BTU per hour. Manual S then selects equipment to match those figures, and Manual D sizes any ductwork.
That process exists precisely because rules of thumb are unreliable, and contractors who skip it are the reason so much residential equipment is oversized. For a plug-in window unit that costs a couple of hundred dollars and can be swapped in an afternoon, shopping from a rule of thumb is entirely reasonable. For a mini-split that will be mounted and refrigerant-charged, a replacement central system, or anything you will live with for fifteen years, the load calculation is worth insisting on and worth paying for — and a contractor who declines to do one, or who sizes from the capacity of the old equipment, is telling you something useful about how the rest of the job will go.
Questions people ask
How many BTU do I need for a 300 square foot room?
Around 6,000 BTU/h as a starting figure, using the common rule of about 20 BTU per hour per square foot with an eight foot ceiling. Realistically it is a range of roughly 5,400 to 6,600 depending on which version of the rule you use, and it moves further with sun exposure, ceiling height, occupancy and how leaky the room is. A 300 square foot room with a wall of west-facing glass under an uninsulated roof genuinely needs more than a shaded north-facing room of the same size, and no square-footage figure distinguishes them. Use the number to decide which two catalogue sizes to look at, not to pick one to the nearest hundred BTU.
Is it bad to get an air conditioner that is too big?
Yes, and it is the more common mistake by a wide margin. An oversized unit cools the air to setpoint quickly and then shuts off, which means it never runs long enough to remove humidity — dehumidification requires sustained runtime with air moving over a cold coil. The result is a room that is cold and clammy rather than comfortable, with noticeable temperature swings and a compressor starting and stopping all afternoon, which shortens its life and uses more energy than steady operation. Undersizing has a much milder failure mode: on the hottest few days of the year the unit runs continuously and holds the room a degree or two above what you set. Given a choice between two sizes, the smaller one is usually right.
How many square feet will a 12,000 BTU air conditioner cool?
On the standard rule of about 20 BTU per square foot, roughly 600 square feet — but that is the answer to a question with several missing inputs. The same 12,000 BTU/h unit is right for a considerably smaller room with a lot of west-facing glass, an uninsulated ceiling above it or a kitchen in it, and right for a larger room that is shaded, well insulated and lightly occupied. Ceiling height matters too, since the rule assumes eight feet. Twelve thousand BTU/h is also one ton, which is the unit central systems are specified in.
Do I need a Manual J calculation for a window air conditioner?
Not usually, and that is a reasonable place to draw the line. A window or portable unit is cheap relative to a professional load calculation, it is not permanently installed, and if the size turns out wrong you can return it or move it to another room. Shopping from a rule-of-thumb range is proportionate. Where the calculation genuinely earns its cost is anything installed and hard to reverse: a mini-split that gets mounted and charged, a replacement central system, or new ductwork. There the sizing decision is locked in for the equipment life, oversizing is common because contractors default to matching the old unit, and many jurisdictions require the calculation for a permit anyway.