Global climate change has led to a gradual intensification of heat island effects and a significant increase in events of extreme high temperature at summer in urbans. How to effectively improve the cooling effect of urban blue-green spaces is an important way to adapt to the climate change and build a living environment with high-quality. The Landsat-8 data and radiative transfer equation was used to retrieve land surface temperature in Kunshan City. The correlation analysis and boosted regression tree （BRT） model of the urban blue-green spatial pattern and the measurement index of summer cooling effect were established constructed at the scale of urban and block to identify the key factors and their importance ranking of urban blue-green spatial patterns that affect summer cooling effects in the Jiangnan water network area and to explore the differences in spatial scale of cooling effect with comparative study. Results showed that there were four indicators with a relative contribution rate of over 10% for the impact of the blue-green spatial pattern on the summer maximum surface temperature at the scale of urban. The decreasing order of impact degree was as follows： index of water body shape > proportion of blue-green space > width of waterfront green space > aggregation index of water body. There were three indicators with a relative contribution rate of over 10% for the impact of the blue-green spatial pattern on the average intensity of cold island at the scale of urban. The decreasing order of impact degree was as follows： width of waterfront green space > index of water body shape > proportion of blue-green space. At the scale of block， there were five indicators with a relative contribution rate of over 10% for the impact of the blue-green spatial pattern on the summer maximum surface temperature at the scale of urban. The decreasing order of impact degree was as follows： index of water aggregation > average patch index of water > ratio of green space > ratio of water surface > index of water body shape. There were two indicators with a relative contribution rate of over 10% for the impact of the blue-green spatial pattern on the average intensity of cold island at the scale of urban. The decreasing order of impact degree was as follows： aggregation index of green space > proportion of blue-green space. It is indicated that the spatial pattern characteristics of urban blue-green space have significant scale effects on the correlation degree and contribution level of summer cooling. The optimization strategies of planning and design were proposed. It will provide a practical reference for creating a comfortable urban living environment at different scales.