Accurate estimation of actual evapotranspiration (ETa) is essential for determining crop water use and improving irrigation management in semiarid agricultural regions such as California’s Central Valley. Thermal infrared (TIR) remote sensing of land surface temperature provides an effective approach for estimating ETa and surface energy fluxes across spatial scales when coupled with energy balance models. In this study, high-resolution TIR imagery acquired from uncrewed aircraft systems (UAS) was used to evaluate the performance of the Two-Source Energy Balance (TSEB) model in almond orchards. Because UAS-based TIR measurements are sensitive to calibration errors and atmospheric effects, a temperature correction procedure was applied prior to model implementation to reduce biases in surface energy flux estimates. The performance of the TSEB-Priestley–Taylor (TSEB-PT) and the two-temperature version of the model (TSEB-2T), which explicitly uses canopy and soil temperatures, was evaluated using corrected TIR imagery. Modeled fluxes were assessed against eddy covariance (EC) measurements and ETa partitioning methods to evaluate the accuracy of evaporation (E), transpiration (T), and ETa. The study specifically investigated (i) whether TIR-corrected UAS imagery improves ETa estimation using TSEB-PT and TSEB-2T, (ii) how hourly evaporation and transpiration from TSEB-2T compare with EC-based partitioning methods, and (iii) how daily ETa estimates derived from the models agree with EC measurements. Results demonstrate that incorporating corrected high-resolution TIR imagery into TSEB-2T improves the reliability of surface energy flux estimates and provides valuable insight into the spatial and temporal dynamics of ETa in almond orchards. These findings highlight the potential of integrating UAS thermal imagery with energy balance modeling to support precision irrigation and water resource management in water-limited agricultural systems.