Surface geometric models provide foundational scenario support for expressing human-machine-environment interaction relationships. This study analyzes the architecture of oblique photogrammetry system, elaborates on key technologies such as image preprocessing, flight strip design, and bundle adjustment, and proposes a complete technical workflow, offering a systematic solution for the reconstruction of surface geometric models. A tripartite accuracy validation framework integrating “data-method-metrics” was designed, where geometric precision was quantitatively evaluated through comparative verification and reprojection analysis. Validation results demonstrate that the model achieves a reprojection error RMSE of 1.03 pixels, with horizontal, vertical, and 3D global accuracy reaching 0.0521 m, 0.0778 m, and 0.0936 m respectively. The study validates the effectiveness of multi-view oblique imagery in high-precision geometric modeling, offering robust theoretical and technical references for environmental applications.