With the increasing scale and complexity of power systems, the rapid and accurate detection of power failures ensures the safe and stable operation of power systems. Traditional fault diagnosis methods rely on manual experience, which has some problems such as slow response and insufficient accuracy. In this study, a comprehensive power fault data processing system based on artificial intelligence technology is proposed. Deep neural network (DNN) model is adopted to classify and detect power fault data, and high-quality data support is provided for model training through data collection, pre-processing, feature extraction and other links. The DNN model has achieved high accuracy in power fault detection, with the classification accuracy reaching 93.4% and fault detection rate 92.0%, and the false positive rate is kept at a low level. It improves the efficiency and accuracy of power fault detection, and provides a reference for the application of artificial intelligence in power system. The research results are of great significance for optimizing the fault handling process of power system and improving power safety.