In this paper, a descriptor Takagi-Sugeno (T-S) fuzzy system is considered, subject to disturbances, sensor and actuator faults, in the presence of unmeasurable premise variables and input saturation, which the later are considered as uncertainties of the system. A robust L2 Proportional-Integral (PI) observer based controller is designed. This observer based controller, can estimate both the system states, the faults, and stabilize closed-loop states trajectories. An integrated robust controller strategy is adopted by a novel structure of non Parallel Distributed Compensation (non-PDC) control law. Moreover, the stability conditions of controller design are expressed in term of Linear Matrix Inequalities (LMIs) constrained optimization problem, by the proposed one-step design procedure. With the help of LMI TOOLBOX in MATLAB, we can easily design the type PI observer for efficient robust state/fault estimation and solve the L2 observer based controller design problem of discrete-time T-S fuzzy systems. The applicability of the proposed scheme is illustrated via a numerical example.