Aero-engine turbine rear casing is an important load-bearing component of aero-engine. Due to its complex working conditions, multiple uncertain factors, it is a key component for aero-engine safety. In order to explore the influence of the uncertainty of input random variables on the failure probability of a turbine rear casing structure, a parametric finite element model was established for the deterministic analysis of the aero-engine intermediate casing. Considering the uncertainty of material properties, geometric parameters and external loads of the aero-engine intermediate casing, limit state functions are constructed for the two most typical failure modes: static strength and stiffness failures. By constructing an adaptive Kriging surrogate model for two failure modes and combining importance sampling method, the failure probability of the intermediate casing structure was predicted. And the uncertainty source of the reliability of the turbine rear casing structure was analyzed by a global sensitivity analysis method based on failure probability. And the importance order of all input random variables was identified, which provides guidance for the reliability design of the turbine rear casing structure.