Aiming at the research of sliding mode control of pump-controlled system, a sliding mode controller based on RBF neural network was designed according to the unknown term f () in the reduced order mathematical model of pump-controlled system, and combined with the sliding mode control algorithm. The simulation model of system was established by MATLAB/Simulink, and then the position command simulation analysis was carried out. The results show that compared with PID controller, the sliding mode controller based on RBF neural network achieves the minimum tracking error. When tracking sinusoidal position signals with 10 Hz frequency and 1 mm amplitude under interference conditions, the sliding mode controller based on RBF neural network has the smallest error. After the interference force is applied, larger tracking errors appear in all controllers, and the sliding mode controller based on RBF neural network can quickly recover the tracking errors. The sliding mode controller of pump control system based on RBF neural network designed has a good tracking accuracy and stronger robustness, which can be widely applied to other mechanical transmission fields.