Green–blue space loss and fragmentation are particularly acute in Chinese cities due to rapid urbanization, large ring-road system and the following city compartments. Therefore, connecting urban green–blue spaces has been recently advocated by central government. This paper revised and applied the recently developed urban green network approach to the case of Xi’an city, China, a city which has been rarely studied before from this perspective. The focus was on connecting fragments of urban green–blue spaces to compact green–blue networks, integrating both social and ecological functions into a fully functioning entity. Landscape metric analysis was added to identify that the main city outside the city core should be a planning priority zone. The Eurasian tree sparrow (Passer montanus), Asiatic toad (Bufo gargarizans) and humans at leisure were selected as three focal species to meet the emerged socio-ecological benefits. Sociotope and biotope maps were drawn up to identify patches with high human recreation and wildlife shelter values and providing crucial network structures. Least-cost-path model was used for identifying potential linkages between patches. This model was based on network structures and cost surface, which measures the theoretical energy cost of travelling between landscape elements. By integrating the potential paths for the selected organisms with density analysis, the updated framework generated three improvement maps for species indicators, and 10 network corridors for establishing green–blue networks at city scale. At neighbourhood scale, one site with habitat and linkage examples illustrated specific measures that could be taken in local practice.