Diabetes mellitus is a common disease among the people. One of the active preventive things to deal with type 2 diabetes mellitus is to do regular exercise, and eat nutritious foods and have adequate nutrition for 1 day. In getting enough calories or energy according to the needs of patients, calculations can be done manually. But if the process is done manually, it will take a long time so that if this is implemented in a health institution, it will be very inefficient given the large number of patients in the queue. This problem can be solved by using an artificial intelligence system using a genetic algorithm that is performed hybridization with simulated annealing. Simulated Annealing can help the genetic algorithm come out of optimum local conditions, due to its nature that can accept solutions that are not better or better than the previous solution. Simulated Annealing was successfully added to help the genetic algorithm out of optimum local conditions, this was indicated by the highest fitness value of 0.998, with the percentage difference between the patient's actual needs of calories by 0.18%, carbohydrate by 0.20%, protein by 0.69% and the last is fat at 0.27%.