Computer interlocking system plays an important role in railway, but its breakdown maintenance cannot stratify requirements of fast developing railway. Prediction model was developed based on mathematical model of components and entire system fault tree analysis, and interlocking system failure probability trend was calculated in the application background of CBTC system. System trend diagram implied that it had three operating condition stages, and corresponding maintenance strategies was linked with critical importance probability of failure events in each stage. The method provided theoretical basis for system prognostic evaluation, maintenance strategy.