Abstract:The roof damage is one of the major cause of hazards in the gob in case of collapse accidents. In the stability analysis of the mined area on the roof with multi-objective evaluation and decision analysis, the extensive use is made of analytic hierarchy process, utility function method, and fuzzy membership function method. Among a variety of objectives, a complex relationship often makes the use of these analysis methods very difficult, especially in determining the weight factors of the objectives, which would affect the application of these methods. In the context of the gob roof accident and the Fault Tree Analysis (FTA), the qualitative analysis based on Bayesian network is transformed by the method of the gob roof fault tree transformation, to establish the gob roof accident Bayesian network model, with the use of packages based on Matlab BNT quantitative calculation, In this way, it is necessary to determine weights and the prior probability and the posterior probability are used in comparison and analysis of the gob to identify the main factors related to the gob roof accidents in the conditional probability, with the construction workers not working in accordance with the design and improper exploitation and support being of the greatest importance, therefore, in the actual construction process, the construction workers must strictly follow the design of mining, to enhance support to ensure the stability of the gob roof.