The dynamic equations of the flexible manipulator are complex and highly nonlinear，the Backstepping design idea is introduced into the controller design and the original complex high-order nonlinear system is decomposed into a low-order simple system．In order to effectively approximate the nonlinear function in the controlled object，it proposes a control method that combines the Ｒadial basis function (ＲBF)neural network with the backstepping control idea．The simulation results show that the neural network backstepping control method based on radial basis function can realize the stability of the flexible manipulator control system．