Input/output (I/O) linearization via state feedback provides a convenient framework for designing controllers for multivariable nonlinear systems. However this approach does not account for parametric uncertainty which may lead to loss of stability and performance degradation. In this paper, a multi-model H-2/H-proportional to approach is utilized to design a robust controller for minimum phase multivariable nonlinear systems that are subject to parametric uncertainty. It is first shown that a state feedback (inner loop) based on nominal parameters introduces nonlinear perturbations to the linear sub-system and results in loss of decoupling. The uncertain system is characterized in a form that provides a framework for robust controller design. Recent results from linear robust control theory are utilized to design an outer-loop controller to account for the parametric uncertainty. This methodology is illustrated via simulation of a regulation problem in a continuous stirred tank reactor.