Different types of underlying land surface cover and various urban spatial morphologies have an important effect on micro-climate parameters. Simulating the spatio-temporal distribution of urban air temperature and analyzing its influencing factors are of great significance to evaluate urban thermal environment and urban planning. In this paper, ENVI-met micro-climate model is used to simulate the spatio-temporal distribution characteristics of surface air temperature in some areas of Nanjing, including Nanjing Middle School I, Eastern Guanghua Street, Xuanwu Lake, and Toutuoling. The distribution of buildings and their heights as well as the underlying land cover information are extracted from the high spatial resolution Geoeye-1 stereoscopic images. The simulated results based on ENVI-met micro-climate model are validated using the measurements collected at Nanjing national meteorological station. It is found that the temporal changes in simulated surface air temperature are consistent with observations. For the spatial distribution characteristics, the surface air temperature distribution in the area of Nanjing Middle School I is similar to that in Eastern Guanghua Street, and the difference in urban spatial morphology results in the air temperature variance within local regions. For the Xuanwu Lake, the surface air temperature decreases outward from the land area in central Xuanwu Lake. The spatial variance of surface air temperature during the daytime is bigger than that in the nighttime due to the complex terrain in Toutuoling area.