The pedagogical method selection is relevant to guarantee adaptability and to personalize each method, while working with intelligent tutorial systems. The objective of this research is just to develop a hybrid system for the pedagogical method selection for an intelligent tutorial system that contributes to knowledge management and strengthens the multiple specific intelligences of each student, in the teaching–learning process. The proposed hybrid system has two components, a genetic algorithm and a neuro fuzzy network, Mamdani Anfis style. The hybrid approach interprets the rules base of a fuzzy system in neural networks terms, where the net simulates a fuzzy inference system of Mamdani type. The learning algorithm works modifying its structure and/or parameters, that is to say, because of neuronal inclusion or exclusion and weight adaptability. The neuro fuzzy network learns in a supervised way, through the Levenberg–Marquardt algorithm. An experiment was applied in order to measure system effectiveness, considering decrease in wrong attempts doing an exercise as a success criterion. The result allows us to confirm that the selection of the pedagogical method generated by the proposal is useful, so it can contribute in a positive way to the programming teaching. This new system enables teaching under in a personalized learning style, taking into account the students characteristics.