Stroke is the leading cause of disability that influences the quality of people's daily life. As such, an effective method is required for post-stroke rehabilitation. Research has shown that a robot is a good rehabilitation alternative where conventional robotic assistive system is encoded program by the robot expertise. The major drawback of this approach is that the lack of voluntary movement of the patient may affect the proficiency of the recovery process. Ideally, the robotic assistive system should recognize the intended movement and assist the patient to perform and make the training exercises more effective for recovery process. The electromyography based robotics assistive technology would enable the stroke patients to control the robot movement, according to the user's own strength of natural movement. This chapter briefly discusses the establishment of mathematical models based on artificial intelligent techniques that maps the surface electromyography (sEMG) signals to estimated joint torque of elbow for robotic assistive system. © 2016 by IGI Global. All rights reserved.