A self-tuning tracking controller for permanent-magnet synchronous motors is disclosed, providing for velocity or position trajectory tracking even when both the electrical and mechanical parameters of the motor, amplifier, and load are initially unknown. A time-scale simplification of a full-order mathematical model of the motor leads to a discrete-time design model that is reduced-order and that evolves in a mechanical time-scale which is substantially slower than the electrical time-scale, permitting implementation of the self-tuning tracking controller with a lower sampling frequency (and at a lower cost) than is typically expected. A piecewise-linear parameterization of the motor torque-angle characteristic functions allows for identification of higher-order harmonics with a degree of accuracy which is selectable by the user, without requiring more computation than traditional single-term sinusoidal descriptions. Excellent performance is achieved, even with very poor initial motor parameter knowledge.