Optimizing green-blue space layouts is crucial in mitigating heat islands in built-up areas, and their cooling effects are influenced by both their own characteristics and the surrounding environments. However, the configuration characteristics of green-blue spaces with high cooling efficiency are still unclear. Additionally, the impact of surroundings on the generation and dispersion of cold sources has not been thoroughly studied. This study establishes a framework to identify the configuration characteristics of green-blue spaces with efficient cooling from four perspectives: occupation, dominance, regularity, and aggregation, and considering building environment, landscape composition, and adjacent influences to control surrounding influences. The Generalized Additive Model was employed to analyze the relationship between these factors and the mean temperature of corresponding units. The results demonstrate that efficient cooling landscape configurations vary among building densities. In low-density groups with an average building height of 3 floors and occupancy of 15 %, scattered small patches can enhance beneficial areas. Conversely, high-density groups with an average building height of 15 floors and occupancy of 32 % require more regular, large-scale patches. These findings offer tailored green-blue space layout strategies for Shanghai and similar large cities to achieve optimal cooling effects while considering limited land resources.