A large number of adjustable – speed drives in industry and emerging applications such as automotive (EV or HEV) require high dynamic performances, robustness against parameter variation and also reliability. parameter detuning and mechanical speed sensor faults lead to a deterioration of the performances and even to instability. therefore condition monitoring is becoming mandatory in those sensitive applications. the objective of this contribution is to study the feasibility of detection and diagnosis of the mechanical speed sensor faults in an induction motor drive. using knowledge of the motor condition, the proposed technique based on fuzzy logic is applied to discriminate load and parameter variations from the speed sensor faults. both simulation and experimental results are presented in terms of accuracy in the detection of speed sensor faults and knowledge extraction feasibility.