In this work, we focus on state and both sensor and actuator fault (SAF) estimations for Takagi-Sugeno (T-S) fuzzy systems in the presence of an Unknown Bounded Disturbance (UBD). To reach this objective, we design a novel observer to estimate simultaneously state and SAF vectors. Indeed, we study the robustness against disturbance by considering the H-infinity performance. For the stability, sufficient conditions are given in term of relaxed Linear Matrix Inequalities (LMIs) where the descriptor formulation is considered. To prove the efficiency of our proposed method, a reduced bioreactor model and simulation results are illustrated.