Abstract Stroke patients with hemiparesis often face significant challenges in accessing consistent and effective rehabilitation therapy due to limitations in mobility and the availability of specialized care. To address this, a new telerehabilitation system is introduced for stroke patients with hemiparesis. It consists of two internetconnected robots for remote wrist rehabilitation. A cloudbased website serves as a centralized interface for patients and doctors, allowing remote control of the system during therapy sessions. To enhance the haptic feedback and stability of the system, a torque/velocity control algorithm is implemented. These control algorithms incorporate damping and spring terms from an impedance control algorithm to synchronize and maintain the positions of both end effectors. Additionally, a wrist evaluation model has been introduced to calculate stiffness and damping coefficients, essential for monitoring patient progress and improving treatment efficacy. The wrist model is evaluated using multiple linear regression method. Experimental testing demonstrates the potential of the proposed approach in enhancing telerehabilitation system performance.