This project investigates the application of model reference adaptive system (MRAS) for the speed sensorless control of an induction motor. The rotor speed can be accurately estimated by employing the closed-loop observer named reactive MRAS. Therefore, this observer eliminates the need of a speed sensor for the control of the motor speed. The method is robust to stator and rotor resistance variations due to change of temperature. The dynamic system equations of the induction machines are formulated, and the motor control system performance is studied. Both scalar voltage-to-frequency (V/f) control and vector field oriented control (FOC) schemes, implemented using digital signal processor (DSP), are investigated. The design of the speed sensorless DSP-based controller is completed. Software packages have been developed to implement the design. An experimental system using the proposed controller has been built. Various tests have been conducted to verify the technical feasibility of the control technique. The experimental results confirm the feasibility of the proposed speed sensorless V/f control scheme using MRAS speed estimator. The designed V/f profile has been tested. Even with step change of the load or that of the command speed, the system can achieve the correct steady state after a short transient operation. The experimental results also confirm the feasibility of the proposed speed sensorless FOC control scheme using MRAS speed estimator. The current regulators meet the design requirements. Both the flux-producing current component and the torque-producing current component can be controlled separately. In the implementation, digital signal processor (DSP) TMS320 FL2407 and voltage source inverter (VSI) Skiip 342GD120-316CTV are employed. The modular strategy is adopted to develop the software package.