Aiming at the risk assessment problem of cascading failures in AC-DC hybrid power grid at different stages of evolution, this paper proposes a cascading failure risk assessment method based on the snow ablation optimization (SAO) algorithm. First, the dominant factors influencing cascading failures in the initial, development and deterioration stages of an AC-DC hybrid power grid are described, and a two-level index system reflecting the evolutionary characteristics of cascading failures is established. Then, a combination weighting strategy for risk assessment indices based on the SAO algorithm is formulated, and the risk assessment process for cascading failures is determined. Finally, a simulation model of an AC-DC hybrid power grid based on an improved IEEE 30-bus system is constructed using the Matlab/Simulink platform. The differences in the weights of the indices between the proposed method and the traditional method are compared, and the effects of different levels of DC converter station blocking on the cascading failures are revealed. The results show that the proposed method can accurately reflect the dominant factors at different stages of evolution of cascading failures of the AC-DC hybrid power grid, and the assessment sensitivity of the high-risk lines can be improved by 12.67% compared with the traditional method, verifying the reasonableness and effectiveness of the proposed method.