Abstract: As the key link of urban rail transit system, the reliability study of metro traction substation is of great significance to ensure the safe and stable operation of the system. In order to evaluate the reliability of the metro traction substation, this paper established reliability model of the typical metro traction substation based on Bayesian Network, calculated the initial failure probability of the system, used the Dynamic Bayesian Network to analyze the reliability of the typical metro traction substation in time dimension, and accurately calculated the curve of system failure probability with time. Finally, the paper used the two-way inference function of Bayesian Network to find the weak link of the system. It has certain significance for the identification, maintenance of key nodes and the optimization of network structure design in substation.