Abstract: The study of fault diagnosis and fault-tolerant control has recently attracted much attention due to the industriaI demands for safety and efficiency. For certain processes, it is important not only to detect(and identify)but also to accommodate any faults quickly. A fault-accommodation controller is presented for a class of nonlinear systems using neural networks approach. An approach for fault accommodation based on an idea of a corrective control law which is activated in the presence of a fault has been developed. Neural networks are used to learn the fault function and provide the corrective action. The resulting closed-loop nonlinear systems are stable after fault accommodation. An example is given to illustrate the accommodation procedure and the effectiveness of the proposed method.