Monthly mean air temperature has influence on evapotranspiration and vegetation growth, and plays an essential part of hydrological process on the land surface. The target of this study is to estimate the monthly mean air temperature over the Asian Continent where meteorological observing stations are quite few. NOAA-11 AVHRR has split window channels at 10.2-11.5, μm and 11.3-12.5 μm. These data are used for the atmospheric correction for estimating not only the sea surface temperature, but also the land surface temperature. Though the split window method also gives good estimations for deriving monthly mean air temperature during April to October in the case study, the appropriate estimation is not obtained through a year. The reasons are considered that the emissivity on the land surface and the relation between the derived brightness temperatures and air temperature change with a season, respectively. Therefore, a new type of the regression equation is proposed in this paper. The equation is consisted of explanatory variables which are split window channels data and a cosine of the zenith distance related to extra-terrestrial radiation. The equation is powerful for estimating the monthly mean air temperature through a year. The regression analyses were carried out at 37 stations for conventinal meteorological observation over the Asian Continent. As a result, many fine regression equations were obtained. Multiple correlation coefficients of derived regression equations were more than 0.95 and standard deviation errors are within 2 degrees Celsius.