An ocean-related geophysical parameter estimation method has been proposed. The method uses a geographic information system (GIS) as a neural network (NN). It was found that the proposed GIS-NN allows geophysical parameter estimations with the most appropriate coefficients for the specified areas and seasons. In comparison to the widely used sea surface temperature (SST) estimation method of multichannel SST (MCSST) with general use of the coefficients for the regression equation using Advanced Very High Resolution Radiometer (AVHRR) bands 4 and 5 from the National Oceanic and Atmospheric Administration (NOAA) satellite, it was found that the proposed method achieves a significant reduction in root mean square error (from 0.315 to 0.245 K). It was also found that the proposed method requires 2.63 times of computation power in comparison with the MCSST method.