An existing algorithm based on the statistical data was refined, and a Python implementation of the improved algorithm was programmed to extract the built-up areas of 656 cities in 1992, 2000, and 2010 based on DMSP_OLS nighttime light data. ArcGIS 10 and Visual FoxPro 6.0 were used to analyze the characteristics of the expansion of the cities. This study selected 2 indices, the compactness of urban external spatial morphology (CUESM) and the speed of urban spatial expansion (SUSE), to analyze the spatial expansion of county- and higher-level cities in China from 3 aspects: the sizes of cities (metropolises, large cities, medium-sized cities, and small cities), large regions (eastern, central, western, and northeastern China), and administrative divisions of China (provinces, autonomous regions, and municipalities). The results are as follows. First, during the same year, the values of CUESM were getting smaller with increasing city, and were larger in central and western China than those in the eastern and northeastern China. During the period of 1992-2010, the values of CUESM became slightly smaller in large cities, almost unchanged in small cities, and got larger in the cities of other sizes. The values of CUESM became smaller in the eastern and western China, but they got larger in the central and northeastern China from 1992 to 2010. Secondly, during the same period, SUSE increased with urban size. From 1992 to 2010, SUSE was high in the eastern China, low in the central and western China, and the lowest in the northeastern China. SUSE was high in Beijing, Chongqing, Shanghai, and Tianjin, low in Guangdong, Jiangsu, Shandong, Zhejiang, and Fujian, and the lowest in the rest of the provinces in China. From 1992 to 2010, SUSE accelerated in all sizes, regions and provinces.