This paper presents a system for the development of new human-machine interfaces focused on static gestures recognition of human hands. The proposal is to aid access to certain objects to the occupant of an intelligent wheelchair in order to facilitate their daily life. The proposed methodology relies on the use of simple computational processes and low-cost hardware. Its development involves a comprehensive approach to the problems of computer vision, based on the steps of the video image capture, image segmentation, feature extraction, pattern recognition and classification. The importance of this work relates to the need to build new models of interaction that allow, in a natural and intuitive way, to simplify the daily life of a disable person.