Abstract: MPC Controllers are good achieving zero steady state error even if the controlled system is affected by disturbances. But MPC strategy by itself was not intended to endure proper operation under faulty conditions. Fault tolerant control consist of control algorithms, fault detection and compensation strategies. In this work a MPC controller is implemented using dynamic matrix control strategy. Also a model based residual generator is developed to detect fault occurrence. Artificial neural networks combined with band pass filter allows fault quantification to compensate the MPC behavior under faulty operation. The implemented compensation schema improved the controller's recovery time significantly.