Abstract: The natural environment influences human society through the medium of "productive forces." In the Anthropocene — an era marked by a dramatic surge in human productivity alongside immense environmental challenges — territorial spatial planning must not only embody the mechanisms of ecological processes, but also respond to the spatial transformations brought about by new production relations involving natural elements. To address the limited attention to natural dynamic evolution and urban-rural interactive processes in current territorial spatial evaluation methods such as the "Double Evaluation," this study proposes a "Natural Portrait" approach as a complementary perspective for territorial spatial cognition. From a "process-driven" perspective, the method deconstructs the historical formation of natural landform patterns, identifies key geographical units, and predicts future evolutionary trends. From a "pattern-response" perspective, it examines the adaptation and transformation of natural landforms under human processes such as urban construction and agricultural production, contributing to an understanding of integrated natural-urban-rural territorial spatial patterns. Through empirical studies of the Guangzhou, Shenzhen, and Western Pearl River Delta metropolitan areas in the Guangdong-Hong Kong-Macao Greater Bay Area (GBA), the research identifies three distinct ecological processes — alluvial deposition, marine deposition, and island-intercepted flotation deposition — each of which has shaped markedly different landform types: concentric plains, headland coastal zones, and circular sand flats. In response to these ecological processes, the three metropolitan areas have developed different territorial spatial patterns — concentric, clustered, and cellular — which continue to evolve toward estuary extension, finger-like expansion, and integrated mosaic development. This method has effectively complemented the existing territorial spatial planning technology chain by enhancing "process cognition" and "unit deconstruction," thereby supporting the development of spatial structures with improved spatiotemporal continuity and more coherent human-land interaction units.