The blue-green spaces, the essential component of green infrastructure, play an important role in promoting the health of urban thermal environment and creating a comfortable microclimate. The typical metropolitan area, the district within the Fifth Ring Expressway in Beijing, was selected as the study area. The basic materials in the study are the Landsat 8 remote sensing image, the land use data and building vector data. We utilized the radiative transfer equation algorithm, K-means clustering algorithm, correlation analysis and other methods to quantify the cooling effect indicators of 1199 blue-green spaces with an area of more than 1 hm2. We analyzed the correlation between the landscape and environment metrics of the blue-green spaces, such as scale, shape complexity, landscape components, environmental components and vegetation coverage, and their cooling effect indicators. Then the blue-green spaces were classified according to the diverse cooling performance. Finally, we discussed the cooling functions, advantages and mechanisms of each category of the blue-green spaces. The results show that:(1) the scale, shape complexity and landscape components of blue-green space have a significant correlation with cooling effect. For example, total area, green space area and water area have the significantly positive correlation with the cooling intensity, cooling area and cooling service and have a significantly negative correlation with the cooling efficiency and cooling gradient. Perimeter of blue-green space has a significantly positive correlation with cooling area and cooling service and have a significantly negative correlation with cooling efficiency and cooling gradient. The ratio of perimeter to area has a significantly negative correlation with the cooling intensity and cooling service and have a significantly positive correlation with the cooling efficiency and the land-surface temperature of the blue-green space. And the shape index has a significantly positive correlation with the cooling area and cooling service. (2) There is also a correlation between the environmental components of the blue-green spaces and their cooling effect. The synergy and promotion of cooling effect occurs between the blue-green spaces. The total area of other blue-green spaces in the cooling area is significantly positively correlated with the cooling intensity, cooling area, cooling service and cooling distance of the blue-green space and is significantly negatively correlated with the land-surface temperature of itself. (3) There are thresholds among the total area, water area, perimeter of blue-green space and its cooling intensity and among the ratio of perimeter to area and cooling efficiency. In order to save the land resources, it is recommended that the indicators above don't exceed the thresholds. (4) Based on the performance of cooling intensity, cooling efficiency, cooling distance and so on, the blue-green spaces are divided into 4 categories. We believe that before the blue-green space planning and construction in metropolitan areas, the government should evaluate cooling demand, understand the cooling conditions of blue-green spaces and reasonably allocate them with suitable cooling functions and advantages. Only in this way can all categories of blue-green spaces exerts the enough cooling effect with the limited land and funds. This study will help the government and planners to understand the current situation, find the problems and development potential of blue-green spaces cooling effect. And it also has some references for the relevant theories and practices.