With the increasing scale and structural complexity of modern power systems, cascading failures have become a central threat to the operational safety of power systems. Such systematic failures may cause catastrophic outages and significant economic losses through cascading propagation mechanisms. In this context, accurately identifying critical links in the power grid has become an important task for power system security defense. For this reason, this study constructs an assessment system for critical links in complex power systems based on the consideration of the impact of cascading failures on the operational risk of large power grids, in order to improve the reliability and anti-interference capability of power grids. In this paper, the ACTIVSg2000 node system is used for testing and analysis to simulate different N-2 and N-3 fault scenarios to assess the disaster loss of the grid, and the simulation results are statistically and analytically analyzed, and the lines are classified to find out the critical lines of the grid, which verifies the correctness of the proposed identification method.