Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) is a multispectral highspatial resolution imaging sensor being developed by Japanese Ministry of International Trade and Industry and one of facility sensors on US National Aeronautics and Spase Administration's EOS-AM 1 platform slated to be launched in June 1998. In this paper, algorithms for water surface temperature (WST) estimation using data from Thermal Infrared Radiomater (TIR), which is one of ASTER subsystems, were investigated through linear regression analyses of simulated ASTER TIR data.<BR>ASTER TIR has five channels in the thermal infrared spectral region (8-12μm) with twelve-bit quantization and expected noise equivalent temperature difference (NEAT) less than 0.3 K. ASTER TIR imagery will be useful for the monitoring of small scale thermal phenomena in coastal regions and lakes because of its high spatial and radiometric resolutions. The capability of multichannel observation of ASTER TIR allows the application of empirical algorithms for water surface estimation such as Split-Window Method.<BR>Brightness temperature values measured by ASTER TIR were calculated using LOWTRAN7 atmospheric transmission and radiance code for various WSTs, atmospheric models, and NEATs. In addition to six atmospheric models originally included in LOWTRAN7 code, twelve atmospheric models were prepared and used in LOWTRAN7 calculation to represent atmospheric characteristics around Japanese Islands. Linear regression analyses of these simulated data indicated that, with NEAT of 0.3 K, expected errors of estimated WST by both Two-Channel and Five-Channel Methods are less than 1.1 K. It was also shown that coefficients of equations for WST estimation depend on atmospheric models used in the regression.