We studied the effects of agricultural development on surface air temperature over Central and Eastern China in the late 20th century by performing four long-term (1980-2000) simulations using the Weather Forecast and Research (WRF) model. The four simulations used exactly the same settings except for the land use and land cover. One simulation referred as control simulation used the satellite-based land use and land cover data of the early 1980s. The other three simulations used the satellite-based land use and land cover data of 2000 only for Northeast, Central and Southeast China. By comparing the seasonal mean temperature of control simulation to ground measurement of 1981-1990, we found that WRF model had a cooling bias of about 2-3 ℃ ; however, WRF model well reproduced the spatial pattern of seasonal mean temperature with the correlation coefficient of 0.91-0.99 (P < 0.001). By comparing sensitive simulations with control simulation, we found that agriculture development resulted in significant changes of local temperature, but with large spatial differences and seasonal differences. In Northeast and Central China, the conversions from forest and grassland to cropland resulted in a cooling effect and the maximum cooling of about 0.41 ℃ occurred in winter. These land use conversions might enhance surface albedo and thereby reduced the surface net solar radiation; as a result, the surface sensible heat decreased and cooling occurred. In Southeast China, the conversion from forest to cropland led to a warming effect with the maximum being about 0.14 ℃ occurring in summer. The deforestation resulted in a flat surface and, thereby, turbulence was suppressed. Then, due to the suppressed turbulence, surface sensible heat was mostly used to heating bottom air rather than emitting into upper air. As a result, the surface climate warming occurred. The abovementioned cooling and warming effects induced by agriculture development could be comparable locally to the climate background changes. However, these cooling and warming effects had very weak modifications on regional mean temperature. This result implicates that the agriculture development-induced cooling and warming effects merely occurred in the locality and had little impacts on regional temperature changes.