This thesis proposes the control of a robotic hand system using surface electromyographic signals (sEMG). The sEMG signals are collected from three different muscle groups of the upper forearm: palmaris longus muscle, extensor digitorum communis muscle, and extensor carpi radialis longus muscle. The objective of this research is to develop a prototype of a robotic prosthesis for people with hand amputation, controlled by an electromyographic interface based on computational intelligence. This thesis covers the following topics: positioning of electrodes to capture the sEMG signals, design of an electromyography muscle interface, preprocessing method, use of techniques of computational intelligence for the interpretation of the sEMG signals, design of the robotic hand, and method used to control the positions of the fingers and of the hand grip force. Here, the wavelet transform is used as a feature extraction method in electromyographic signals, and a multi-layer neural network as a pattern classification method. The proposed model obtained satisfactory results, recognizing 90.5 per cent of the positions for 6 different hand patterns, 94.3 per cent for 5, and 96.25 per cent for 4 positions.