The invention discloses a detection method and device of the fault degree of a high-voltage circuit breaker spring. The method comprises steps of taking an acquired circuit breaker journey signal as an objective, and processing and analyzing characteristics of mechanical vibration signals, thereby achieving fault diagnosis of the circuit breaker spring; by acquiring journey signals of the circuit breaker spring in the operation process, extracting the average speed of the movement of a contact and the average standard variance of the fluctuating quantity from the journey signals so as to form characteristic vectors; based on a constructed fault classifier, carrying out fault classification; and based on a fitting function where the pre-compression change quantity of the spring is taken as the independent variable and the average standard variance of the fluctuating quantity is taken as the dependent variable, carrying out judgment of the fault degree. According to the invention, performance of a high-voltage circuit breaker opening/closing spring can be monitored in real time only by acquiring the journey signals; faults can be timely diagnosed without additional vibration sensors; the data is simple to process; the economic performance and the practicability are high; and the method and the device have excellent application prospects.