The overall purpose of our work is to take advantage of Thermal Infra-Red (TIR) imagery to estimate landscape evapotranspiration fluxes over agricultural areas, relying on two approaches of increasing complexity and input data needs: a Surface Energy Balance (SEB) model, TSEB, used directly at the landscape scale with TIR forcing, and the aggregation of a Soil-Vegetation-Atmosphere Transfer (SVAT) model, SEtHyS, run at high resolution (≃100 m) and constrained by assimilation of TIR data. Within this preliminary study, models skills are compared thanks to large in situ database covering different crops, stress and climate conditions. Domains of validity are assessed and the possible loss of performance resulting from inaccurate but realistic inputs (forcing and model parameters) due to scaling effects are quantified. The in situ data set came from 3 experiments carried out in southern France and in Morocco. On average, models provide half-hourly averaged estimations of latent heat flux ( LE ) with a RMSE of around 55 W m −2 for TSEB and 47 W m −2 for SEtHyS, and estimations of sensible heat flux ( H ) with a RMSE of around 29 W m −2 for TSEB and 38 W m −2 for SEtHyS. TSEB has been shown to be more flexible and requires one single set of parameters but lead to low performances on rising vegetation and stressed conditions. An in-depth study on the Priestley-Taylor key parameter highlights its marked diurnal cycle and the need to adjust its value to improve flux partition between sensible and latent heat fluxes (1.5 and 1.25 for south-western France and Morocco, respectively). Optimal values of 1.8 to 2 were hilighted under cloudy conditions, which is of particular interest with the emergence of low altitude drone acquisition. SEtHyS is valid in more cases while it required a finer parameters tuning and a better knowledge of surface and vegetation. This study participates to lay the ground for exploring the complementarities between instantaneous and continuous dynamic evapotranspiration mapping monitored with TIR data.