Abstract: In view of the shortcomings of neural network algorithm and fuzzy control algorithm in the application of maximum power point tracking of photovoltaic power generation, An adaptive neural-fuzzy controller is proposed, It is a combination of fuzzy control and neural network control, And it can produce fuzzy rules and adjust membership functions automatically based on data without experience of experts. The simulation model is built by using MATLAB/SIMULINK software and MATLAB toolbox. Under the same environmental conditions, adaptive neural fuzzy control method, fuzzy control method and perturbation and observation method, they are conducted to simulation experiments Respectively. The results of simulation show that the controller has good dynamic performance and steady state accuracy in photovoltaic MPPT, which improves the working efficiency.