In this study, a sliding mode controller with an adaptive fuzzy impedance model is proposed to improve the human–robot cooperation and promote lower limb rehabilitation performance using a hybrid robot including an orthosis and a programmable plate, designed and built at the University of Guilan. Orthosis controls the patient in the path of the joints, and the plate is considered instead of a treadmill to prevent impact on the soles of the feet. Since the patient’s cooperation to regain motor skills is of particular importance in the training process, the proposed controller tracks the path of the joints while considering the patient will. The controller consists of an adaptive impedance control whose coefficients are adjusted using a fuzzy logic. In fact, the impedance model defines the relationship between the robot and subject. Using the model, the subject force is converted to a virtual displacement and, then, added to the tracking error, then a sliding mode controller tracks the modified path. Moreover, the Lyapunov function is used to guaranty the stability and the sign function which depends on the patient cooperation. To evaluate the performance of the proposed method, the control model is implemented on a healthy person and, then, is compared with the sliding mode controller and constant impedance sliding mode controller. The results show that this method while following the path of joints is able to consider the subject cooperation and overcome nonlinear terms while chattering is low.