Abstract : This works deals with the stabilization of uncertain and/or disturbed nonlinear systems represented by fuzzy Takagi-Sugeno models. First, some results based on a quadratic candidate Lyapunov function have been proposed in terms of LMIs (Linear Matrix Inequalities) and the conservatism of such approaches has been discussed. To reduce the conservatism, new approaches based on a non quadratic Lyapunov function and non PDC (Parallel Distributed Compensation) control law have been proposed. Then, one other source of conservatism has been studied. Indeed, the classical way to express the closed-loop dynamic leads to introduce crossing-terms in the set of LMIs to be solved. Therefore, to overcome this source of conservatism, the descriptor redundancy propriety has been used to rewrite the closed-loop dynamic. This allows decoupling the system matrices from those of state feedback gain matrices. Furthermore, the redundancy propriety has been employed to cope with the wellknown and difficult problems in terms of LMI formulation relating to the robust static and dynamic output feedback controller design for uncertain and / or disturbed Takagi-Sugeno systems.