In order to avoid the calculation error of discrete T-S fuzzy fault tree,to clearly express the variation trend of system reliability with time,the combination of the continuous-time T-S dynamic fault tree based on impulse function constraint and dynamic Bayesian network could not only reflect the dynamic and static description ability of the continuous-time T-S fuzzy dynamic fault tree,but also give full play to the advantages of two-way inference modeling and describable dynamic of Bayesian network.The Bayesian network model was built in GeNIe2.0 software,and the posterior probability was derived.Then,the importance degree of the synthesized dynamic and static modules was calculated.Finally,taking the EMU air supply system as an example,the reliability analysis was carried out,and the variation curve of system reliability with time was simulated in the software.The correctness and rationality of the fuzzy dynamic Bayesian network method were verified.The results show that the safety valve and the check valve are the weak links of the system,so the frequency of inspection and maintenance should be increased.The other parts can be reasonably arranged for equipment maintenance frequency according to the results shown in the diagram.