This study aims evaluating regional distribution, and diurnal and seasonal variations of surface heat budget over a rice paddy, using limited number of satellite infrared images, meteorological data, as well as a surface heat budget model. Data used in this study were NOAA-AVHRR thermal-infrared images (channel 4) for retrieving surface temperature, surface and upper meteorological data, and incoming solar radiation. In order to evaluate surface heat budget, number of satellite data is not sufficient because cloud conditions and satellite paths are not appropriate. To overcome this difficulty, not only satellite data, but also a numerical heat budget model were used. This model is linear, prognostic, and consists of two layers (vegetation canopy and underground soil), which was modified from a surface heat budget model developed by Matsushima and Kondo (1995) for a bare soil surface. Number of required satellite image was only one for a diurnal variation, and an appropriate time of satellite passby was near noon. Estimation errors were 20-30 W m^-2 for both sensible and latent heat fluxes, compared with observations by the Bowen ratio method on half-hourly basis. Implementing solar radiation retrieved from satellite data to the above method and application to other types of vegetation are future issues.