Considering the model uncertainty and disturbances of FreeFloating space Robot, an Neural Network based robust control scheme is proposed aim to solve the problem of trajectory tracking in joint space for free floating space robot. Using RBF neural network to approximate the nonlinear function of controller, robust controller is used to compress the approximation error and external disturbance. The presented control scheme can stabilize the system and guarantee the gain from disturbance to tracking error is lower than the given index. Simulation results show the feasibility of the proposed control scheme.