Abstract: Since the drawbacks of conventional sliding mode control in maximin power point tracking(MPPT), such as slow convergence speed and remarkable chattering, a nonlinear backstepping integral sliding mode MPPT control of PV system based on RBF neural network is proposed. Firstly, RBF neural network is used to predict the output voltage of photovoltaic cells under various conditions. Secondly, a nonlinear integral sliding mode surface is designed to improve the steady-state error and the problems of large overshoot. Finally, a new exponential approach law is designed to accelerate the convergence speed while effectively weakening the high-frequency chattering of the system. The reachability and stability of the nonlinear backstepping integral sliding mode control are analyzed by Lyapunov functions, the simulation experiments are conducted under static, dynamic and shading conditions. The simulation test results show that under the working conditions where the temperature and irradiation change, the nonlinear backstepping integral sliding mode control based on RBF neural network can track the maximum power point of the photovoltaic system quickly and accurately under various meteorological conditions compared with the traditional sliding mode control, and has strong robustness.