The estimation of surface energy fluxes using remotely sensed data requires the combination of data from several sources, including land use, vegetation cover and surface temperature. Land use and vegetation cover were obtained from visible and near infrared (VNIR) data, while the state variable, surface temperature, was obtained from thermal infrared (TIR) data. An approach to combine these data with an energy balance model was studied as part of the 1997 Southern Great Plains Experiment (SGP97). Toward this end, VNIR and TIR images for 2 July 1997 were analyzed over the El Reno, Oklahoma (OK) site using data from the Thermal Infrared Multispectral Scanner (TIMS) and Thematic Mapper Simulator (TMS) airborne instruments. Intensive ground measurements constrained leaf area indices, canopy height and surface meteorological inputs required by the model. The observed brightness temperatures, when corrected for atmospheric effects using MODTRAN, nearby radiosoundings and the temperature-emissivity separation (TES) algorithm, were mostly within VC of ground based temperatures. The resulting surface temperatures were used in a two-source model that considers the heat flux and temperature contributions from the soil and vegetation. The heat flux predictions on average agree within 50 W m−2 of tower-based observations.