This paper presents a new protection scheme for synchronous generator using an artificial neural network. The proposed scheme is able to discriminate internal and external faults of synchronous generator, consists of feature extraction using eigenvalues and fault classification using artificial neural networks. The performance of algorithm was tested using artificial neural networks and simulation results show the proposed scheme is able to identify and classify all types of faults of synchronous generator. The proposed approach has the ability to detect the generator winding faults which are close to neutral.