The main task of air rudder electromechanical servo system is to receive control commands and drive the load to swing at the commanded angle,which has an increasingly wide range of applications in various aerospace vehicles.However,the unknown disturbances in the operation process bring great challenges to the high-precision control of electromechanical servo systems.To address this problem,an air rudder servo system control strategy was proposed based on disturbance compensation.The model analysis of the air servo motor servo system was carried out.The radial-basis-function-based neural network state observer was designed to replace the unmeasurable rudder angle with the estimated value for feedback control.Finally,the finite-time convergence condition was analyzed by applying the Lyapunov method.The simulation results show that the proposed control strategy reduces the steady-state error of the air rudder servo system by more than 97% compared to the conventional motor angle feedback.