【Objective】To address the issues of imprecise modeling and high computational cost in structural damage detection using the finite element method (FEM), this study proposes a structural health monitoring (SHM) technique that combines the spectral element method (SEM) with the Northern Goshawk Optimization (NGO) algorithm.【Method】Firstly, the spectral element method was used to establish the frequency response function of the structure, which was then applied to construct the objective function for damage localization and detection. This approach divided the damage detection problem into two stages, reducing the optimization dimension and complexity. Secondly, NGO algorithm was introduced to optimize and solve the objective function. Finally, planar truss structure and ASCE Benchmark Structure were used as case studies to compare the damage detection performance of NGO, Particle Swarm Optimization (PSO), and Ant Lion Optimization (ALO) algorithms under various damage cases.【Result】The results show that for low-dimensional and simple structures, NGO, PSO, and ALO algorithms all exhibit good solving capabilities. However, for high-dimensional and large complex structures, NGO demonstrates superior damage detection capability and robustness compared to PSO and ALO.【Conclusion】The improved method enhances the accuracy of numerical modeling in damage detection.