The actuator plays a very important role in the flight control system of UAVs, so it is necessary to diagnose the fault in real time.Analysis was made to the working principle of the actuator of UAVs, based on which the analytical model of the actuator was established.With this model, we made real-time status monitoring and fault diagnosis to the actuator.Then, a fault diagnosis algorithm based on residual was designed by combining threshold analysis with Improved SPRT.The fault features were extracted from residuals to detect and determine the working status of the actuator.Finally, a mathematical model of actuators fault was established, and fault models were injected to the actuator to verify it.The result shows that this scheme can identify the common faults of the actuator rapidly and effectively.