Habitat fragmentation leads to a sharp decline in biodiversity, affecting habitat ecosystem services. A rational and effective ecological network construction can restore the connection between broken habitat patches in cities. The construction and simulation of the existing urban ecological network are mostly based on two-dimensional data information, and the impact of the multi-ecological niche represented by the three-dimensional structure information of vegetation on the habitat quality is not well considered, so the function effectiveness of the ecological network remains to be verified. This paper quantified the three-dimensional green space vegetation structure based on the airborne LiDAR and high-resolution remote sensing image data in the main urban area of Nanjing city. We extracted the ecological source area and built the landscape resistance surface in two-dimensional and three-dimensional scenarios, and identified the ecological corridor under the two scenarios based on least-cost path method. Meanwhile we identified the ecological key nodes in the corridors that need to be protected according to the circuit theory, then analyzed the impact of using two-dimensional and three-dimensional ecological spatial information on the results of ecological network construction. The results show that: 1) there are 11 different ecological sources extracted under different scenarios within the study area. Compared with the traditional two-dimensional scenario, using three-dimensional vegetation parameters can more accurately identify ecological sources with rich vegetation structures. 2) There are 137 and 129 ecological corridors constructed in the two-dimensional and three-dimensional scenarios, and the average unit resistance of each corridor is 18.2 and 24.0, respectively. The traditional two-dimensional information simulation overestimated the landscape connectivity of the study area. 3) The spatial distribution of key ecological nodes in different scenarios is similar, mainly concentrated in the northwest, middle and southeast of the study area, and more often appear in the source patches with close distance. Based on the above results, this paper proposes some targeted strategies to guide the construction and optimization of the ecological network pattern in the main urban area of Nanjing city.