Abstract: Background and Objective: In this study, a fifth-order Discrete-time Extended Kalman Filter (DEKF) approach is introduced for on-line estimation of speed, torque and rotor position in a Permanent Magnet Synchronous Motors (PMSM). Methodology: In order to verify the effectiveness of the estimation, the DEKF scheme has been implemented and validated in real-time on a test bench, which is composed of a synchronous machine, a powder brake completed by current and voltage sensors. Results: The satisfying experimental results carried out on a 1.6 kW PMSM demonstrate the excellent performance of the proposed DEKF against measurement noises and perturbations. Conclusion: The interesting results obtained on the PMSM show that this estimation method can be to use for the reconstitution of the speed and position in the industrial applications.