Probabilistic networks can be a useful tool for reliability modelling and risk analysis in the operation of technical facilities. In the article, the authors present the method of using probabilistic networks in the analysis of threats to the elements of the electric bus power supply system with a hydrogen fuel cell. Both the use of probabilistic networks as a tool supporting decision-making processes and the use of the latest achievements in the field of non-emission propulsion indicate a new area of research. The article presents the construction of a hydrogen bus with particular emphasis on the hydrogen storage and transmission installation. Next one specific case is considered where there is a leak in the hydrogen system with the simultaneous appearance of an ignition initiator in the form of a short circuit in the electrical system. Authors considered Fault Tree-Event Tree method in context of Bayesian networks technology. The conducted research has shown that there is a real risk of hydrogen explosion in a hydrogen fuel cell-powered bus and that it can be accurately calculated.