Power system protection is important for service reliability and quality assurance. Various faults may occur due to natural and artificial calamity. To reduce the outage duration and promptly restore power services, fault section estimate has to be done effectively with appeared fault alarms. Dispatchers could study the changed statuses of primary/back-up relays and circuit breakers to identify the fault section and fault types. It is difficult to process too many alarms under various conditions in a large power system. Single fault, multiple faults, single and multiple faults could coexist with the failed operation of relays and circuit breakers, or with the erroneous data communication. Dispatchers need more time to process the many uncertainties before identifying the fault. This thesis presents the use of artificial intelligence for fault section detection in substation with neural networks. Probabilistic Neural Networks (PNN) are proposed for fault detection system in substation. The proposed methodology will use primary/back-up information of protective relays and circuit breakers to detect the fault sections involving single fault, multiple faults, or fault with the failure operation of the relays and circuit breakers. This paper also presents a fuzzy theory-based method to identify fault types. It is derived to improve the inadequacy of making decisions by selecting a fixed threshold value and has the capability of non-deterministic decision making with a prior knowledge of uncertainties in fault location, fault resistance and the a size of loads. The proposed approach has been tested on a typical taipower system with accurate results.