Objective Rapid urbanization since the 1970s has degraded ecosystem services globally, posing significant challenges for metropolitan areas. In highly urbanized regions such as the Beijing-Tianjin-Hebei region, the Yangtze River Delta, and the Guangdong-Hong Kong-Macao Greater Bay Area, three interrelated challenges persist: fragmented spatial concepts and policy integration, regional management disconnections at administrative boundaries, and spatial conflicts between biodiversity protection and public recreation. While existing research explores open spaces and green infrastructure separately, few establish a coherent logic linking conceptual integration, spatial identification, and governance tools. This study addresses this gap by developing an integrative framework that transforms open spaces into green infrastructure networks through spatial policies to enhance ecosystem service functions.

Methods This study adopts a two-stage analytical framework of “international mechanism reference—local framework exploration.” First, systematic policy analysis of The London Plan identifies three core pathways for operationalizing open spaces: legal designation (statutory protection of high-value areas), planning prescription (strategic spatial reservation through plans), and development control (regulatory requirements for new developments). Building on this conceptual grounding, the study employs quantitative spatial analysis methods tailored to the Greater Bay Area context. Landscape character assessment establishes an open space baseline by extracting ecological, agricultural, and eligible urban park landscapes. Integrating natural thresholds with urban land distribution delineates a tripartite structure of outer bay barrier, green belt, and inner bay core zones. Strategic open spaces are identified, including regional green belts along the outer-inner transition, metropolitan green belts between sub-metropolitan areas, and metropolitan open spaces based on statutory plans. The InVEST model generates habitat quality maps classified into core, important, and general areas. Morphological spatial pattern analysis extracts forest cores larger than 1 km², landscape connectivity indices identify ecological sources, and circuit theory-based linkage mapping extracts potential corridors and bottlenecks. Public open spaces are classified into five tiers by size and service radius to identify deficiency areas. At the local scale in Panyu District, regional outputs are overlaid with local planning contexts to propose gain-oriented control indicators tailored to strategic locations.

Results The empirical analysis yields three main findings. First, the open space baseline reveals a fundamental contrast: the outer bay dominated by natural mountains and agricultural landscapes, the inner bay core by high-density urban areas interspersed with rural river networks—where ecological resources are scarce yet human activities intensive. The metropolitan boundary demarcates an inner area of approximately 23,000 km². Within this structure, two green belt types are identified: regional green belts along the outer-inner transition containing inter-city sprawl and safeguarding outer bay ecological integrity; and metropolitan green belts between the three sub-metropolitan areas preventing urban coalescence. This yields a tripartite spatial structure of outer bay barrier, the green belt, and inner bay core zones, with open space area ratios of approximately 7∶1∶2—supporting differentiated governance: ecological conservation in the outer bay, strategic regulation in the green belt and targeted mending in the inner core. Second, potential designated areas reveal critical gaps across protection, connectivity, and service dimensions. Biodiversity assessment shows 81.88% of high-value cores outside the ecological protection redline. Ecological network analysis identifies 419 sources (19,486.93 km²) and 898 corridors (5,410.74 km), revealing a source vacuum within the inner metropolitan area except for isolated low hills, while the dense river network provides natural corridor carriers. Public open space accessibility analysis shows deficiency areas account for 47.81% of the inner metropolitan area, with 17.80% in dense urban landscapes—indicating a disconnect between green space availability and service accessibility. Third, in Panyu District, the framework demonstrates local applicability. The district contains metropolitan green belts (11,000.12 hm², 21.03%), metropolitan open spaces (1,974.56 hm², 3.78%), and biodiversity conservation areas at regional (1,694.42 hm², 3.23%) and local (10,535.3 hm², 20.14%) levels. Critical corridors along the Shawan Waterway and Shiji River are identified. Public open space deficiency reaches 56.80% district-wide, with eastern areas experiencing service deficits despite abundant open space due to productive agricultural land use.

Conclusion Theoretically, this study clarifies the conceptual distinction and complementarity between “open space” as spatial carrier and “green infrastructure” as functional network. Drawing on London’s experience, it extracts three core pathways—legal designation, planning prescription, and development control—through which green infrastructure networks are operationalized by protecting existing open spaces or requiring new developments to provide them. Through systematic comparison, the study translates international experience into localized insights, identifying key areas for optimizing China’s spatial planning instruments to better align with its institutional context and support metropolitan ecosystem services. Empirically, the research establishes an open space baseline under a unified concept in the Greater Bay Area, constructs a regional spatial structure of outer bay barrier, strategic green belt, and inner bay core zones, and identifies three strategic area types: biodiversity priority areas, regional ecological corridors, and public open space deficiency areas. In Panyu District, it demonstrates how optimizing the three tool types can address metropolitan complexities and achieve enhancement in ecosystem service functions. Together, these findings offer an integrated planning approach connecting conceptual unification, spatial identification, and governance tools—a pathway beyond fragmented green space management. Future research could deepen gain-oriented indicator design, develop standardized methodologies, test the framework across diverse metropolitan contexts, and explore legal and fiscal mechanisms for implementing innovative tools within China’s institutional framework.