The health units continually receive patients with clinical pictures that point to a stroke, however, these health institutions do not always have at their disposal specialists in neurology and computerized axial tomographs that allow timely differentiation of the type of stroke. In this case, the speed of diagnosis is decisive in the conduct to be followed by medical and paramedical personnel, in order to reduce the risks that lead to deterioration of health or a fatal outcome for the patient. In support of the action protocols for stroke, which have been described internationally in medicine, a group of researchers decided to identify a classification method to develop a computer system to differentiate ischemic and hemorrhagic strokes using intelligent algorithms in the event of inconclusive cases. The basis of the proposal is based on the rules. NNge algorithm, a supervised classification algorithm, by means of which 1165 production rules were defined, which have more than one antecedent and a single consequent for each rule. This consequent is the type of stroke or cerebrovascular accident presented by the patient.