Abstract: Nowadays, a new type of electromagnetic repulsion mechanism has been applied in high-voltage vacuum circuit breakers. In order to diagnose the types of faults occurring in electromagnetic repulsion mechanisms, a fault diagnosis method based on the improved S-transform and support vector machines is proposed in this paper. Firstly, the electromagnetic repulsive force mechanism is used to obtain vibration signals at two different positions of the high-voltage vacuum circuit breaker during operation. Then, an improved S-transform is used for time-frequency analysis of the vibration signal, and feature quantities are extracted based on the normalized energy entropy; the principal component analysis is used to reduce the dimension of the feature vector. Finally, the feature quantities are input into the SVM training model for fault classification, and compared with wavelet packet decomposition. The results show that the proposed fault diagnosis method has good performance and can quickly and accurately identify the faults occurring in the electromagnetic repulsive force mechanism.