In order to effectively measure the flight speed of the aircraft, the thermal measurement and pressure measurement used to calculate the true airspeed measurement are studied. Combined with the relationship among pressure altitude, calibrated airspeed, mach, total temperature parameter, true airspeed of flight data recorder, fit true airspeed and carry out residual comparison through least squares support vector regression (LS-SVR). According to the feature of LS-SVR, calculate different true airspeed at deferment time. The results show that the LS-SVR method has high precision and strong generalization ability, which can effectively monitor and diagnose the performance of total temperature sensor performance drift and fault, and has application value.