High voltage circuit breakers mostly use spring operating mechanisms, which are prone to faults due to their complex structure. Strengthening the research on latent fault diagnosis of circuit breaker operating mechanisms and detecting latent faults in advance is beneficial for the reliable operation of circuit breakers. It is of great significance for the safety and stability of power systems. This paper focuses on a vacuum circuit breaker spring operating mechanism and sets up a vibration acceleration test system. Using examples such as loose buffer fixing bolts and loose hanging plate fixing bolts, we extract fault diagnosis feature quantities based on the wavelet packet band energy decomposition algorithm, determine the diagnostic threshold of latent faults using empirical formulas, and predict the development trend of fault features using the least squares fitting method. The results show that this method can effectively identify and diagnose latent faults in circuit breakers, providing reliable technical support for the maintenance of high voltage circuit breakers.