This paper describes the development of tracking and discriminate methodology for human eye movements. This methodology is the first stage of a research project oriented for the development of a human-machine interface for disability patients. The purpose of project is help to solve some common problems of this people, such as the impossibility to communicate and to interact with the environment. The methodology uses digital image processing for eye tracking and computational intelligence methods for eye movement classifying. The methodology is software-based develop for real time operation. Several tests for a particular patient and for others were made. Results of test show the efficacy of the methodology for a particular patient. However, the develop needs to improve his robustness for work with several patients.