Abstract: Atmospheric transmittance is an important parameter in the thermal infrared remote sensing. A multi-variable lookup table of atmospheric transmittance which includes atmospheric model, aerosol model, water vapor content, visibility and view zenith angle was constructed based on the radiation transfer model MODTRAN, effect of different parameters on thermal infrared atmospheric transmittance spectrums was analyzed, the key variables of atmospheric transmittance were determined by the analysis of variance. According to different types of aerosol model, the multi-variable linear regression models of atmospheric transmittance models were deduced based on the water vapor content, visibility and view zenith angle for common thermal infrared sensor channels, which will solve the problem on calculating accurately the atmospheric transmittance for the thermal infrared remote sensing by satellite.