According to the World Health Organization, approximately 70% of adults in Mexico are overweight or obese, determining factors in the development of diabetes mellitus type 2. In addition, according to the National Institute of Public Health, 10.3% of those over 20 years old suffer from diabetes. To facilitate decision or classification tasks when treating a patient, experts develop systems based on fuzzy logic, however, this design is not usually infallible, so it is common to optimize them to improve their performance. The present work shows the results of a comparison between the efficiency in predicting the risk of suffering from type 2 diabetes established by the FINDRISC test and an own design fuzzy system optimized by the Simulated Annealing heuristic for 295 patients from Acapulco, Mexico. The comparison shows that the fuzzy system obtains the same sensitivity, but higher specificity values and positive and negative predictive values with general improvement in the confidence intervals, concluding that using the proposed system as an aid in the prevention of type 2 diabetes is viable and yields results attached to the reality of the patients.