In this paper, we treat the problem of the sensor faults detection for induction motor. The false measures or the disconnections of these sensors produce severe transient errors in the estimator for the control loop and involve ominous effects on the performances of the motor. Therefore, a fuzzy observer is designed to estimate both states and sensor faults. Firstly, the Takagi-Sugeno (T-S) fuzzy model is employed to approximate the induction motor in the synchronous d-q frame rotating with electromagnetic field-oriented. Secondly, a descriptor observer is synthesized where the sensor faults are considered as an auxiliary variable. The observer parameters are obtained by solving a set of Linear Matrix Inequalities (LMIs). Finally, simulation and experimental results are given to illustrate the effectiveness of the proposed approach.