A method to retrieve land surface temperature (LST) is proposed in the thermal infrared (8-12μm) from MODIS and ASTER data. Firstly, spectral mixture analysis was conducted between MODIS and ASTER VNIR channels and the result showed that the overall accuracy was about 5%. Secondly, the band average emissivities are calculated using spectral response function of MODIS sensor and published spectral data of terrestrial materials in wide rages of atmospheric and surface temperature conditions. Thirdly, LST was calculated using generalized split-window algorithm using the coefficients from regression analysis of radiative transfer simulations proposed by Z. Wan [Wan, 1996] . Comprehensive validation and error analysis has been made to evaluate the performance of the new LST algorithm using NASA/MODIS LST product (MOD11) and AMEDAS data. The maximum error in retrieved LST values were 1K. Results show that the new approach used with MODIS and ASTER data offers an improved retrieval of LST.