Abstract : Fuzzy modeling and identification of nonlinear systems are addressed in the first part of this research work. Unstationary, perturbed and multiple inputs and single output systems (MSO) are considered. Based on the learning input-output variables (off-line obtained) a modeling and identification method is applied on a Bio-climate system, which is an experimental greenhouse from the Universite du Sud Toulon Var - France (USTV). A neuro fuzzy system is then synthesized, having a fuzzy rule structure proposed by Takagi-Sugeno-Kang (TS) where the premises are identified via the fuzzy C-Meaas clustering algorithm. The multi-objective approach is implemented by the local and global learning in the consequents parameters. By means of the TS - like rule base, a new approach to model SISO systems is introduced. The fuzzy Gustafson-Kessel clustering algorithm is applied in order to identify the premises of the rule base, and the consequents of the rule base are cubic polynomials. In the second part of this research work, by some Linear Matrix Inequalities (LMI) based on the greenhouse fuzzy model obtained, a solution of a stable controller is obtained. This controller acts on the internal temperature in order to regulate the VPD parameter in a suitable range for the crop plant.