Abstract: It is important for pipeline safety operation and management to improve the accuracy of the quantitative risk analysis of natural gas pipe failure. Based on various information, Bayes network bears strong modeling and analysis capability to process the questions in system polymorphic, nondeterministic and dynamic characters described with current probability safety assessment method. The factors of natural gas pipeline damages are introduced. By means of utilizing the fault tree algorithm to transfer to Bayes network, having established the long transmission pipeline Bayes network quantitative analysis model, the pipeline failure probability calculation method is also found. On the basis, the case study and failure analysis on some long-distance pipeline with HUGIN Bayes software are performed, which realizes the modification in event polymorphic and logistic nondeterministic problem when using the fault tree to analyze.