Time-varying failure rate is one of important indicators of protection devices’ state assessment. With the development of on-line monitoring technology, especially the construction of new generation smart substations makes the temperature parameter of protection devices can be observed. This paper does research on time-varying failure rate of protection devices based on the temperature indicator. The paper combines Arrhenius model with Weibull distribution model and focuses on the time-varying aging failure rate under considering the operating temperature of the modules. Then the paper integrates the aging failure rate with the occasional failure rate of each module with the series model. At last, the paper establishes the time-varying failure rate model of a certain provincial grid’s protection devices in the example, taking a device as an example, analyses the impact of operating temperature difference and replacing internal module on failure rate briefly. The result shows that temperature has significant influence on failure rate of protection device, and the influence will extend as time goes on.