Model reference adaptive controllers with Minimal Control Synthesis are effective control algorithms to guarantee asymptotic convergence of the tracking error to zero not only for disturbance-free uncertain linear systems, but also for highly nonlinear plants with unknown parameters, unmodeled dynamics and subject to perturbations. However, an apparent drift in adaptive gains may occasionally arise, which can eventually lead to closed-loop instability. In this article, we address this key issue for discrete-time systems under L 2 disturbances using a parameter projection algorithm. A consistent proof of stability of all the closed-loop signals is provided, while tracking error is shown to asymptotically converge to zero. We also show the applicability of the adaptive algorithm for digitally controlled continuous-time plants. The proposed algorithm is numerically validated taking into account a discrete-time LTI system subject to parameter uncertainty, parameter variations and L 2 disturbances. Finally, as a possible engineering application of this novel adaptive strategy, the control of a highly nonlinear electromechanical actuator is considered.