The “observation minus reanalysis” (OMR) method is used to estimate the impact of land surface forcings on surface temperature by computing the difference of trends between the observations and reanalysis data set. The basis of this method is that if observed surface temperature, moisture, and wind over land are not used in reanalysis, the reanalysis data set should not be sensitive to urbanization and other land use change. In this study, the OMR method is used to estimate the impacts of land surface change on surface temperature trends during the period from 1979 to 2008 in eastern China, and the difference time series of extreme temperature indices between the observations and the reanalysis data sets is also analyzed. The OMR trends of annual cold (warm) nights exhibit generally decrease (increase), it means that land surface change is likely to enhance the decreasing (increasing) trends of annual cold (warm) nights. The decreasing (increasing) OMR trends of annual cold (warm) days is found in semiarid region in northern China, while the increasing (decreasing) of cold (warm) days occurs in eastern agricultural area. For eastern China as a whole, the land surface change impact may explain about one third of the observed increase for the annual warm nights and nearly half of the observed decrease for the annual cold nights, and the impacts on the annual cold days and warm days are relatively small. The land surface change may reduce the diurnal temperature range significantly. The impact of land surface forcings on extreme temperature demonstrates obviously annual variation.