Based on the data envelopment analysis, taking capital, land, and labor as the input factors, and non-agricultural GDP as the output factor, the paper quantified the urban economic efficiency of cities at the county level and above in China, and identified its main factors by the panel binary selection model. The results show that urban technical efficiency is higher in eastern China than that in western China, and it has a certain spatial spillover globally. But the spillover effect of the urban scale efficiency is much smaller, and its high scale efficiency is mainly close to that of the small towns around large cities, especially in the Bohai Rim region, the Yangtze River Delta region and the Pearl River Delta region. Secondly, based on the Malmquist index, China's urban technological efficiency changes and urban economic efficiency changes have obvious local spatial characteristics, that is, economically developing regions are better than the developed regions, and small cities are better than large cities. And the cities with high urban scale efficiency changes are mainly concentrated in the major urbanization areas such as the Bohai Rim region, the Pearl River Delta region and the Yangtze River Delta region, and large cities have faster growth than small cities. Thirdly, there is obvious spatial heterogeneity in China's urban economic efficiency fluctuations. And urbanization level, industrial pollution discharge, government scale, science, and technology investment, fiscal policy and social development are the main positive factors affecting urban economic efficiency fluctuations, while real estate development investment has no obvious influence. Besides, with the continuous growth of urban economic scale and the decreasing scale of urban construction land, urban economic efficiency will gradually rise. When it reaches a certain threshold, urban economic efficiency will decline, and the large-scale urban sprawl at this stage does not bring about a substantial increase in urban economic efficiency.