Abstract: Due to the effect of the environmental loads and the defects of its own, the casing in shale gas wells may suffer deformation failure in the process of service, which will bring serious impact on the normal production of shale gas wells. In order to control the risk induced by the deformation failure of the casing in shale gas wells, its causes and consequences were identified by means of event tree and event sequence diagram. Then, the scenario model for the deformation failure of the casing in shale gas wells was established, and event tree and event sequence models were transformed into the Bayes network. The inference relations between events were presented by using conditional probability table, and the Bayes network model for the deformation failure of the casing in shale gas wells was established. Failure diagnosis and analysis and probability learning were carried out, and the most possible cause for the deformation failure of the casing was identified. Threatened accident was introduced and the dynamic occurrence probability of each failure scenario was assessed. Finally, model application study was carried out with shale gas well A in the eastern Sichuan Basin as the example. And the most possible cause combination for the deformation failure of the casing was figured out, and the dynamic occurrence probability of casing deformation failure and its consequence was calculated. It is indicated that they are in better accordance with the safety situation of the casing in gas wells in the field of engineering. The research results can provide the theoretical support for the risk control and decision making on the deformation failure of casing in shale gas wells.