Aiming at the problems of backlash，input saturation and unmeasured state information in the design process of electromechanical actuator controller，an output feedback control method based on adaptive neural network state observer is proposed.In order to describe the influence of backlash on system dynamics，a fourth-order servo system model with backlash is constructed by introducing an approximate dead zone function.A state observer based on adaptive neural network is designed for the unmeasurable state to realize the reconstruction of system state under the condition of model uncertainty.Under the framework of backstepping method，an output feedback controller is constructed by using the hyperbolic tangent Lyapunov function and state observer output.For the possible input saturation，an auxiliary filtering system is introduced to compensate the influence of input saturation.The boundedness of error signals in the closed-loop system is proved based on Lyapunov theory.The effectiveness of the proposed control method is verified by conducting multiple sets of simulation experiments.The results demonstrate the superior performance of the proposed method in suppressing the effect of backlash nonlinearityt on the system performance，compensating for input saturation，and achieving the actuator accurate tracking control with unmeasured state information.