<p>This work presents a robust output-feedback linearization framework for uncertain nonlinear systems that, as a natural consequence, yields the dynamic structure of a Proportional-Integral (PI) controller with an implicit anti-reset windup (ARW) mechanism, while preserving the fundamental properties of nonlinear control. The approach employs a reduced-order high-gain observer to estimate lumped uncertainties by introducing an additional state that captures unmodeled dynamics, external disturbances, control implementation errors, and parametric uncertainties, which are compensated through the control input. Based on Lyapunov theory, the resulting closed-loop system is shown to be globally practically stable and globally asymptotically stable as uncertainties vanish. The proposed formulation does not require an exact plant model. It allows straightforward tuning based solely on an estimate of the plant’s open-loop time constant and a lower bound on the system nonlinearities. Moreover, the resulting control law requires the same number of arithmetic operations as a classical PI controller with ARW, making it suitable for practical implementation. As an example application, the proposed framework is used to regulate the inductor current in a Buck converter with parasitic resistances. Simulation results under parametric uncertainty, load-step disturbances, sampling effects, quantization, and actuator saturation demonstrate improved robustness compared with a robust linear PI+ARW benchmark. This benchmark is constructed by augmenting the linearized Buck-converter model with integral action, tuning the resulting state-feedback gains through a Linear Quadratic Regulator (LQR) design with passivity-based weighting matrices, and adding an explicit anti-reset windup mechanism.</p>