In this paper, a discrete model reference adaptive controller is designed and implemented to let the performance of the field-oriented induction motor drives be insensitive to the parameter changes. Only the information of the reference model and the plant output are required for the control. No online explicit parameter identification is required. Hence, the proposed controller is easy to implement practically. For designing the proposed adaptive controller, the dynamic model of the drive system is estimated from the sampled input/output data using the stochastic modeling technique. Theoretical basis of the adaptive control is derived and simulation is made. Then the hardware of the drive system and the microprocessor-based adaptive controller are implemented. Some experimental results are given to demonstrate the effectiveness of the proposed controller.