Abstract: High permeability renewable energy interconnection taking power electronic equipment as interface has become a prominent feature of future power distribution networks. Renewable energy possesses randomness and intermittence,and power electronic equipment as grid-connection interfaces also will lead to the deterioration of power quality. In order to improve power quality,a neural terminal sliding mode control for active power filter is proposed in this paper. Firstly,using fractional order scheme and sliding mode theory,a fractional-order terminal sliding mode control（FOTSMC）is designed to ensure the finite-time error convergence,and the boundary layer technique is utilized to eliminate the chattering problem. Moreover,a model-free control scheme is designed using selforganizing fuzzy neural network（SOFNN）to better handle uncertainties. The SOFNN controller is designed for learning FOTSMC,which not only solves the chattering problem fundamentally,but also inherits the finite time convergence performance of FOTSMC and satisfies the stability control performance in the framework of Lyapunov theory. Simulation and experimental results demonstrate that the proposed control methods can effectively solve the harmonic problem in the renewable energy generation system.