In view of the complicated flexible manipulator position control problem, puts forward an adaptive sliding mode control method, which combined with pole assignment technology. The variable structure sliding mode control is applied to the rigid motion and elastic vibration suppression of flexible manipulator and the pole assignment is used to set the sliding mode surface pole to obtain good dynamic response. By means of RBF network adaptive learning system upper bound of the uncertainties, the neural network output is used for adaptive correcting the switch gain of control law. Instance simulation results show that the proposed scheme is capable of manipulator position control, and meanwhile it can effectively restrain the elastic vibration of flexible arm and is robust to uncertain parameters.