Abstract: This paper proposes an intelligent predictive control scheme for power electronic drive systems under complex operating conditions, based on neural network adaptive gain. The scheme introduces an RBF neural network compensator for online learning of system nonlinearities and intelligent adjustment of control gains. By integrating predictive control with neural networks, the system′s predictive accuracy and performance are enhanced. MATLAB/Simulink simulation validation shows that the scheme has good adaptability and robustness, effectively improving the system′s dynamic response and disturbance rejection capabilities.