An active fault compensation control law is developed for a class of nonlinear systems to guarantee closed-loop stability in the presence of faults, based oil a neural network representation of the fault dynamics. Changes in the systems due to faults are modeled as unknown nonlinear functions. The closed-loop stability of the robust fault compensation scheme is established in the Lyapunov's sense. The nonlinear fault function is represented by a neural network, then all adaptive corrective control law is formulated to ensure system stability. The main contributions presented are the design of the fault compensation and corrective control law of nonlinear systems with unmatched uncertainties and the stability analysis of the closed-loop systems in the presence of fault modeling errors. Applications of the proposed design indicate that the fault compensation control law is effective for a nonlinear fermentation process. (C) 2008 American Institute of Chemical Engineers.