Due to the influence of evaporation on land-surface temperature, thermal remote sensing data provide valuable information regarding the surface moisture status. The Atmosphere-Land Exchange Inverse (ALEXI) model uses the morning surface temperature rise, as measured from a geostationary satellite platform, to deduce surface energy and water fluxes at 5–10 km resolution over the continental United States. Recent improvements to the ALEXI model are described. Like most thermal remote sensing models, ALEXI is constrained to work under clear-sky conditions when the surface is visible to the satellite sensor, often leaving large gaps in the model output record. An algorithm for estimating fluxes during cloudy intervals is presented, defining a moisture stress function relating the fraction of potential evapotranspiration obtained from the model on clear days to estimates of the available water fraction in the soil surface layer and root zone. On cloudy days, this stress function is inverted to predict the soil and canopy fluxes. The method is evaluated using flux measurements representative at the watershed scale acquired in central Iowa with a dense flux tower network during the Soil Moisture Experiment of 2002 (SMEX02). The gap-filling algorithm reproduces observed fluxes with reasonable accuracy, yielding ∼20% errors in ET at the hourly timescale, and 15% errors at daily timesteps. In addition, modeled soil moisture shows reasonable response to major precipitation events. This algorithm is generic enough that it can easily be applied to other thermal energy balance models. With gap-filling, the ALEXI model can estimate hourly surface fluxes at every grid cell in the U.S. modeling domain in near real-time. A companion paper presents a climatological evaluation of ALEXI-derived evapotranspiration and moisture stress fields for the years 2002–2004.