Ice surface temperature (IST) controls the rate of sea ice growth and the heat exchange between the atmosphere and ocean. In this study, high-resolution IST using the Advanced Spaceborne Thermal Emission and Reflection radiometer (ASTER) thermal infrared region (TIR) images was retrieved to observe the thermal change of coastal sea ice. The regression coefficients of the multi-channel equation using ASTER brightness temperatures ( B T ) and MODIS ISTs were derived. MODIS IST products (MOD29) were used as an in situ temperature substitute. The ASTER IST using five channels from band 10 ( B T 10 ) to band 14 ( B T 14 ) showed an RMSE of 0.746 K for the validation images on the Alaskan coast. The uncertainty of the two-channel ( B T 13 and B T 14 ) ASTER IST was 0.497 K, which was better than that of the five-channel. We thus concluded that the two-channel equation using ASTER B T 13 and B T 14 was an optimal model for the surface temperature retrieval of coastal sea ice. The two-channel ASTER IST showed similar accuracy at higher latitudes than in Alaska. Therefore, ASTER-derived IST with 90 m spatial resolution can be used to observe small-scale thermal variations on the sea ice surface along the Arctic coast.