This paper proposes a scheme for position control of a mechatronic servo-system based on gain-scheduled PIDs (known as NPID in related literature) and adaptive approximation of uncertainties. NPID and adaptive approximation are combined in a constructive way to inherit advantages and overcome limitations of the single methods. In particular, NPID is used to stabilize the nominal plant, and its gains are scheduled so as to minimize the effects of friction and other uncertainties. Adaptive approximation is used to compensate further effects of nonlinearities, and obtain a better overall tracking accuracy. A report on an extensive experimental investigation is provided to illustrate the practical advantages of the proposed scheme.