Abstract: Researching the application of Bayesian network to the system reliability assessment for the limits of combination method and state method in reliability assessment.The method is able not only to compute the system reliability indices but also to analyze easily the effects of a certain components on the system reliability to recognize the vulnerable positions of the system. A method of constructing Bayesian networks with the ideology of fault tree is proposed to assess system reliability and the validity of the method proposed is verfied by exemplification. The results of two examples prove the validity and superiority of the method in the application of the system reliability assessment. Contrasting and analyzing the conditional failure probability of Bayesian network and the importance measurements usually used in system reliability engineering getting the result that the inference arithmetic is more suitable to recognize the vulnerable positions of system.

Key words: system reliability assessment, Bayesian networks, fault tree, importance, inference