Legs are vital parts of the human body and are used for movement and standing. Many people around the world might lose their legs or lose the ability to walk normally due to accidents or diseases. Although there are many walking assistant devices available clinically and commercially, such as prosthetics and orthotics, few are developed to be automated and safe. This situation motivated me to study this issue and develop an EMG- based automated knee-orthosis device. In the proposed method, an EMG signal was extracted from the motor neurons connected to the knee muscle while in movement. The extracted signal was then used to drive a motor. The EMG signal was extracted using an EMG sensor placed on the upper knee muscle while the muscle was moving and the angle of the leg was recorded using an accelerometer placed on the lower limb below the knee. Both EMG and accelerometer signals were then sent to a PC using MATLAB through a DAQ (Humsoft MF 624) for processing and control. In the first stage, these signals were pre-processed using a low- pass filter. Then, the relationship between the leg angle and EMG signal was estimated using the MATLAB's System Identification Toolbox. This nonlinear relationship was modeled using a Hammerstein–Wiener model. This mapping was used to obtain the reference signal used in a closed-loop trajectory tracking control system. The proposed control system works as follows: 1) The EMG signal is measured from the upper knee muscle and transformed to an angle signal using the identified nonlinear transform. 2) The transformed signal is low-pass filtered and used as a reference input of the closed-loop system to control a motor with a load. To assess the quality of the proposed EMG-based control system, the angle output of the servomotor was compared to the angle's signal measured from the real moving leg. This control system approach showed good matching between the two signals. Based on the obtained results, I expect that the proposed approach to be a better alternative to the widely used methods based on pattern recognition methods. The same approach was used to control a DC motor. To enhance the tracking capability of the proposed method, a PID controller was used for controlling the DC motor position. The performance of the proposed control method further compared to one of the standard EMG-based control method, namely the threshold-based method. The implementation results of both methods were compared by assessing the error between the real and desired leg angles.