In this paper, neural network control is presented for a rehabilitation robot with unknown system dynamics. To deal with the system uncertainties and improve the system robustness, adaptive neural networks are used to approximate the unknown model of the robot and adapt interactions between the robot and the patient. Both full state feedback control and output feedback control are considered in this paper. With the proposed control, uniform ultimate boundedness of the closed loop system is achieved in the context of Lyapunov’s stability theory and its associated techniques. The state of the system is proven to converge to a small neighborhood of zero by appropriately choosing design parameters. Extensive simulations for a rehabilitation robot with constraints are carried out to illustrate the effectiveness of the proposed control.