24 plots composed of commercial districts, school areas, residential areas and parks in Beijing were selected to extract land surface temperatures by using remote sensing and GIS method. Land cover types were acquired, and the pattern of ecological green space was evaluated by a landscape software. The authors investigated the relationship between land surface temperatures of various plots and underlying ground structures. The results show that the water region and the ecological green space have the cooling function, while the concrete surface can heat the land surface. The proportions of the above three types strongly affect the average land surface temperature of the explored plots. The effect of selected ecological green space indices on temperature is in decreasing order of ‘green space ratio’, ‘division’, ‘average area of patch’, ‘cohesion index’, ‘shape index’, and ‘fractal dimension’. There is a positive correlation between ‘resolution’ and temperature, and a negative correlation between temperature and other quantities. Among all kinds of the selected plots, the descending order of their temperatures is from commercial districts through school areas and residential areas to parks. The temperature distributions of parks are extremely scattered, while those of commercial districts are the most concentrated.