Maintaining habitat connectivity is one of the major challenges facing global biodiversity conservation. The establishment and optimization of habitat networks provide an offensive landscape strategy, but its effectiveness depends on habitat distribution patterns and species response in a heterogeneous landscape. Taking Egretta garzetta as the target species in Nanjing, the study established and optimized the habitat network based on grouping analysis and LSCorridors simulation. Firstly, by integrating habitat suitability evaluation, disturbance evaluation, and MaxEnt model simulation, the comprehensive identification of species habitat was realized. Then, using the landscape connectivity index and kernel density analysis, the grouping pattern of habitat distribution was analyzed. Finally, two kinds of optimization corridors (i.e., the added corridor and the improved corridor) were defined and identified based on the least-cost paths method, while their multiple random paths and their spatial ranges were thereby simulated by LSCorridor software from the perspective of the variation of species perception and adaptation. The results showed that: there were 428 habitat patches in Nanjing, with a total area of 31525.75 hm², mainly forest and waterbody, with evident fragmentation and local aggregation situation. Nine habitat groups, including landscape patches of the Yangtze River, were concentratedly distributed on both sides of the river, but ecological connections among them were relatively weak. Within an optimal distance threshold of 4.5 km, 907 least-cost paths with a total length of 1548.59 km presented an averagely short and spatially compact status and were mainly located in the interior of each habitat group, while there was a lack of paths connecting the habitat groups. For optimization, eight added corridors and 10 improved corridors were extracted to be further stimulated by LSCorridors, showing the spatial deviations from the least-cost paths. The simulation outputs were considered to be more in line with reality, thus reducing the inefficiency of conservation and promoting the formulation of multi-type and differentiated optimization measures. The study results can provide a decision-making reference for habitat network protection and key ecological corridor planning practices for urban biodiversity conservation.