Abstract: High voltage ac SF6 gas insulation switch equipment defect identification depends on the preset threshold to determine the existence of fault, the actual operation of the equipment may be affected by a variety of factors, makes the fault characteristics deviate from the preset threshold, resulting in low identification accuracy, to study the high voltage ac SF₆ gas insulation switch equipment defect identification method based on RBF neural network. The high-frequency sensor is used to monitor the electrical signal in real time to obtain the operation status information of the equipment. Advanced signal processing technology is used to extract the time-frequency domain features of the electrical signals. On the basis of feature extraction, a defect identification model based on RBF neural network is constructed to identify the defect type and severity of GIS equipment. The experimental results show that the simulation waveform is highly consistent with the actual condition under the simulated grounding fault condition, and the recognition confidence is up to 0.995, which is significantly better than the comparison method and has high practical application value.