At an early stage of developments, brake performances are usually obtained from a simulation using a brake dynamometer. However the brake parts on the dynamometer will not operate at exactly the same temperature as that fitted to a road vehicle, because the cooling factors are different. Also the load on the brakes on the dynamometer is different from the actual load. The effects of temperature and load on brake performances were investigated and it was found that over a considerably wide range the frictional coefficients of brake materials are only minimally affected by the cooling factors and load. Brake performances calculated from the experimental data obtained from dynamometer tests using different line pressures and velocities as parameters show good coincidence with the actual road test results, so the simulations can be used to effectively estimate brake performance at planning stage of the vehicle.