This paper describes the application of an identification algorithm clustering type Gustafson-Kessel nonlinear
dynamical system. From input-output data the algorithm generates fuzzy models of Takagi-Sugeno. This type of modeling is applied
to a non-linear numerical model. The non-linear input / output model of the system is decomposed in several described by
membership functions and fuzzy rule-based local linear systems. The results are presented and prospects for future work.