Abstract: Based on the analysis of principle and hardware architecture of computerized interlocking system, this paper proposed a novel method of life prediction by using fault data. The fault tree model of the system was built, from the calculation of the minimum cut set of the fault tree, the key factors that affecting the service life were achieved; and the board-level fault data was classified in accordance with the Exponential and Weibull distribution statistical rules, which made the life prediction model more practical. Finally, taken a widely applied A/B computerized interlocking system as an example, the simulation result showed that the proposed method was feasible. The work in this paper could provide reference for the maintenance management of similar signaling equipment.