A new approach is described for fuzzy modeling of control systems. It is shown that conventional concepts and criteria of stability and controllability of a control system can be extended to a fuzzy setting. More precisely, we establish a fuzzy dynamical model for nonlinear control system, that can be consistently applied to all linear multiinput/multioutput (MIMO) systems. This modeling is a fairly general representation of fuzzy control systems, and satisfies some natural extensions of the criteria such as stability and controllability, that conventional control experts use to evaluate a control system.<<ETX>>