Power system reliability assessment is an important part of power system analysis, and plays a vital role in the overall planning, design and operation of the system. In this paper, the Bayesian network model is applied to the reliability evaluation of the power generation and transmission system. Firstly, the first-order fault events enumerated is pretreated. Secondly, based on the minimum cut set, a Bayesian network model considering the joint faults of components is constructed, and the minimum load shedding of the power system is obtained through the AC optimization model. Moreover, the conventional reliability index is calculated, and the weak link of the system reliability is identified. Finally, the proposed methodology is applied to the IEEE RBTS 6 Bus Test System, but representative, the numerical study shows its feasibility in larger systems.