The aim of this paper is to present a new framework for designing robust controllers for uncertain nonlinear systems described by Takagi–Sugeno fuzzy descriptor systems and subject to input and state constraints. For this, a novel structure of modified parallel distributed compensation control laws is proposed. The controller parameters can be obtained through the numerical resolution of an optimization problem with linear matrix inequality constraints derived from Lyapunov's stability theory and the choice of an original Lyapunov function. This ensures the stabilization of the closed-loop system and the respect of the state and input constraints regionally. The technique is illustrated through three numerical examples.