The advancement of electronics enables the development of more complex functional prosthetics with more capabilities and functionality. The prosthetics currently developed do not leave the laboratory environment due to their complexity or the requirement to operate in conjunction with laboratory computing equipment. Therefore, it is essential to develop prosthetics that can operate outside of the laboratory environment while maintaining maximum functionality. Consequently, an electronic control system is developed that can be embedded in a functional prosthetic hand, which is then trained in real-time through electromyography (EMG) using a classification algorithm based on machine learning. The system allows the prosthetic hand to assume six different positions and be controlled by electromyography, recognizing flexion and extension movements of the wrist. This results in a control system that can be fully embedded in the prosthetic hand, with an appropriate weight and size to maintain the aesthetics and functionality of the prosthesis.