Context: Although many studies have considered urban structure when investigating urban ecological networks, few have considered the three-dimensional (3D) structure of buildings as well as urban green spaces. Airborne LiDAR datasets provide an opportunity to quantify 3D structure and evaluate 3D metrics in connectivity mapping. Objectives: We examined an urban ecological network using the 3D structure of both green spaces and buildings. Methods: Using breeding-season bird species observations and airborne LiDAR data collected, we assessed the influence of 3D structural variables on species diversity. We used correlation analyses to determine if vertical distribution, volume, area, and height of both buildings and vegetation were related to bird species diversity. Then we conducted circuit theory-based current flow betweenness centrality (CFBC) analysis using the LiDAR-derived structural variables. Results: We found that the volumes of buildings and 8–10 m vegetation heights were both highly correlated with species richness per unit area. There were significant differences between 2D and 3D connectivity analysis using LiDAR-derived variables among urban forest patches, boulevards, and apartment complexes. Within urban forest patches and parks, 3D CFBC represented canopy structural characteristics well, by showing high variance in spatial distributions. Conclusions: The 3D CFBC results indicated that adjacent high-rise buildings, dense apartment complexes, and densely urbanized areas were isolated, as characterized by low centrality values, but that vegetation planted in open spaces between buildings could improve connectivity by linking isolated areas to core areas. Our research highlights the importance of considering 3D structure in planning and managing urban ecological connectivity.