The analysis of the models is carried out starting from experimental data of a multivariable system MISO (Many Input Single Output). The models’ implementation was made using fuzzy logic. In fuzzy logic, the cluster technique was used to decrease the number of rules to use in the identification. This technique is opposed to the conventional method which requires a considerable number of fuzzy inference rules to approach the model. In the consequence of fuzzy model, different techniques are used to implement Takagi-Sugeno type rules. By other hand, we implemented the Neuro-fuzzy modeling methods, which let represent the non-linear system and at the same time a system with some learning degree using different topologies. By comparison the goodness of each method is obtained.