To address the issues of low efficiency and strong subjectivity in traditional geometric morphology assessment of steel structures, an automated assessment method for geometric morphology of steel structures by incorporating laser point clouds was proposed. Grounded in classical algorithms including denoising and clustering techniques, the proposed method achieved automatic segmentation of target point cloud data. To overcome limitations in precise segmentation of point clouds data of complex steel structure, a manufacturing deviation calculation method based on local spatial relationships of point clouds was developed, which was then applied to detect point cloud data of complex steel structure while providing recommended parameter configuration ranges. Through local projection analysis and triangular mesh topological relationship evaluation, the method completed the automatic surface defect identification and quantitative analysis, resolving the long-standing challenge of quantitative metric deficiency in visual inspections. Laboratory results demonstrate that the proposed method effectively evaluates geometric morphologies of steel structures with complex configurations, verifying its reliability and practical utility.