Addressing the challenge of insufficient accuracy in building regional cultural heritage 3D models using single-image modeling techniques, a method for optimizing regional ancient architectural three-dimensional models through the fusion of laser and imagery was proposed. Initially, imagery data of the target area was acquired through drone-based cross-flight aerial photography combined with close-range photography. Subsequently, laser scanning data was obtained using a 3D laser scanner to cover blind spots from drone aerial photography. Then, the laser scanning data was fused and registered with the imagery data to generate a complete point cloud model of the target area, which is then used to reconstruct a refined 3D model of the regional cultural heritage. Finally, the superiority of the proposed method was validated through error analysis of the fused heterogeneous data and comparative analysis of accuracy and texture completeness with single-image modeling results. The results indicate that the fused heterogeneous data achieves high fitting accuracy, and the resulting regional ancient architectural 3D model exhibits high accuracy and texture completeness, thereby providing valuable technical reference for the detailed modeling of ancient architectural complexes within the target scope and holding broad application prospects in enhancing the digital archive storage of ancient architectural three-dimensional models.