Looking at max power dissipation is important, but sometimes isn't specified since it depends on thermal mounting. Calculating can be harder as well.
The first rule of MOSFETs is to never use them in the linear region - but with a logic level MOSFET and digital on/off control there is no risk there (there are linear FETs - rare and $$).
If you are just switching something on or off infrequently, you can generally ignore the on/off region of the switching time and just look at the MOSFET Rds(on). When fully switched on, a MOSFET acts like a resistor with the specified Rds(on) value. Note that Rds(on) can change quite a bit with gate voltage, so look for curves. Even logic level FETs may only have the lowest on resistance when the gate is above 5V.
If you're using them in a rapid PWM fashion, you need to count the time the MOSFET is transitioning between states (where its in a linear region, and power dissipation can be extremely high for short periods of time), which can vary how hard its driven, and if there is any ringing or other oscillations on the gate. This is hard to measure without proper tools and can be a bear to accurately simulate, so my rule of thumb is to healthily derate current/power by at least 50% when considering PWM applications for a FET.