Urban parks composed mostly of vegetation and water bodies can effectively mitigate the urban heat island effect. Many studies have investigated the cooling effects of urban parks; however, little attention has been given to park landscape structure. Based on landscape metrics, this study has explored the influences of the park landscape structure on its inner thermal environment, taking heavily urbanized Beijing Municipality in China as the study area. Three indices, including the percentage of landscape(PLAND), landscape shape index(LSI) and aggregation index(AI), were used to measure the composition and configuration characteristics of the landscape components inside the parks. The indices were calculated for five landscape types being interpreted from Quickbird images. Urban thermal conditions were measured using the land surface temperature(LST) derived from Landsat TM images. The results showed that the park LST had a negative relationship with the park size, but no significant relationship was found with park shape. For the park’s interior landscape, however, the configuration and composition characteristics of the landscape components inside the park explained 70% of the park LST variance. The area percentage of water bodies and the aggregation index of woodland were identified as the key influencing characteristics. In addition, when the composition and configuration characteristics of the park landscape components were separately considered, the configuration characteristics(LSI and AI) explained approximately 54% of the variance in park LST, which was comparable with that explained by the composition characteristics(PLAND). Thus, this study suggested that an effective and practical way for urban cooling park design is the optimization of spatial configuration of landscape components inside the park. Urban parks composed mostly of vegetation and water bodies can effectively mitigate the cooling effects of urban parks; however, little attention has been given to park landscape structure. Based on landscape metrics, this study has explored the influences of the park landscape on its inner thermal environment, taking heavily urbanized Beijing Municipality in China as the study area. Three indices, including the percentage of landscape (PLAND), landscape shape index (LSI) and aggregation index (AI), were used to measure the composition and configuration characteristics of the landscape components inside the parks. The indices were calculated for five landscape types being interpreted from Quickbird images. Urban thermal conditions were measured using the land surface temperature (LST) derived from Landsat ™ images. The results showed that the park LST had a negative relationship with the park size, but no significant r For the park’s interior landscape, however, the configuration and composition characteristics of the landscape components inside the park explained 70% of the park LST variance. The area percentage of water bodies and the aggregation index of woodland were identified as the key influencing characteristics. In addition, when the composition and configuration characteristics of the park landscape components were separately considered, the configuration characteristics (LSI and AI) explained approximately 54% of the variance in park LST, which was comparable with that explained by the composition characteristics (PLAND). Thus, this study suggested that an effective and practical way for urban cooling park design is the optimization of spatial configuration of landscape components inside the park.