Abstract: Network security inspection indicators and evaluation are currently the mainstream supervision and inspection methods in railways, and are important measures to promote the formulation of security strategy decisions. This paper analyzed the applicability and limitations of fault tree analysis and neural network methods, proposed an evaluation method of railway network security inspection based on Analytic Hierarchy Process, set network security inspection indicators and indicator weights in different levels, and conducted inspection case validation. The research results can provide technical reference for the subsequent annual inspection of railway network security.