Abstract: The systems in nuclear power plants have complex sequence failure behaviors due to redundant design, and the traditional static fault tree(SFT) in risk assessment cannot accurately simulate the sequential failure behaviors of the systems in nuclear power plants. In this paper, an accident risk analysis framework based on event tree+Dynamic Fault Tree(DFT) is proposed to solve this problem, and the main steam line break(SLB) accident is used as an example to carry out the analysis. First, an event tree model of the SLB accident and the DFT model of the related system are established; Then, the system fault tree is divided into DFT modules and SFT modules, and the result of the DFT modules are replaced with super events to participate in subsequent calculations; Finally, the cut set method is adopted to calculate and analyze SLB accident risk, and results are compared with the traditional SFT method under the same conditions. The results show that (1) Compared with the SFT method, the result calculated by the proposed method is closer to the real system failure scenario, and the conservativeness of the result is relaxed; (2) The proposed method can accurately simulate the SLB accident, which is beneficial for operational risk management and maintenance optimization in the future.