At present, most of the industries focus in the field of control engineering. It concerns to monitor the power system in order to detect the occurrence of fault in the system, identify the type of fault and also its location. It has been mainly done to prevent the system from the faults. A vital component in power system network is transformer, where the modern power transformers are subjected to different types of fault. Due to this reason, the continuity of the flow of current will get affected. Therefore, various fault diagnostic approaches are adopted to identify the faults in the power transformer. The power transformers have been emerged as an integrated part of the power system. Any fault occurred in the transformer may cause a severe outage, so that continuous monitoring and diagnosis of its operation is required. The performance of this proposed work is estimated with some performance metrics such as, accuracy, error rate, standard deviation, training time and testing time. During the initial phase, the proposed Optimization techniques are compared with the existing techniques.This research work had been aimed to develop an efficient fault diagnosis strategy.