Ecological networks (ENs) are increasingly used to address habitat fragmentation caused by accelerating urbanization for biodiversity conservation. The selection of target species significantly influences EN outcomes. However, most studies rely on a limited number of high-priority targets and therefore risk overlooking the needs of coexisting taxa. To address this limitation, we developed a trait-based virtual species framework for regional mammal conservation planning in Zhejiang, China. We clustered all 99 terrestrial mammal species into eight virtual species based on habitat preference, dispersal capacity, and minimum area requirements, and used these groups to guide habitat identification, scenario comparison, and EN construction. Specifically, suitable habitats were identified for each virtual species, and a nested set of conservation scenarios (CSs) was generated according to increasing degrees of habitat overlap. The scenarios were comparatively evaluated using conservation effectiveness (CE), landscape connectivity (LC), and conservation cost (CC), and the highest-ranked scenario was selected for subsequent EN construction. We constructed separate ENs for each virtual species and integrated them to form the final multispecies EN. Key findings include: (1) suitable habitats for Zhejiang mammals extend beyond pristine forests to semi-natural and human-modified mosaics; (2) within the evaluated scenario set, CS4 retains high CE (0.94) and LC (0.84) while reducing CC by 49.71% compared to the full-union scenario CS1; and (3) the resulting EN highlights the importance of numerous short corridors for dispersal-limited taxa, whereas longer corridors for medium- and high-dispersal groups may require mitigation at major infrastructure crossings to maintain connectivity. Overall, this virtual-species framework provides a transferable planning-support workflow for regional multispecies EN planning under limited species-specific data, supporting a shift from single-species surrogacy toward trait-based multispecies planning and enabling structured comparison of trade-offs between ecological performance and relative conservation burden.