A two-source model for deriving surface energy fluxes and their soil and canopy components was evaluated using multi-angle airborne observations. In the original formulation (TSEB1), a single temperature observation, Priestley–Taylor parameterization and the vegetation fraction are used to derive the component fluxes. When temperature observations are made from different angles, soil and canopy temperatures can be extracted directly. Two dual angle model versions are compared versus TSEB1: one incorporating the Priestley–Taylor parameterization (TSEB2I) and one using the component temperatures directly (TSEB2D), for which data from airborne campaigns over an agricultural area in Spain are used. Validation of TSEB1 versus ground measurements showed RMSD values of 28 and 10 Wm−2 for sensible and latent heat fluxes, respectively. Reasonable agreement between TSEB1 and TSEB2I was found, but a rather low correlation between TSEB1 and TSEB2D was observed. The TSEB2D estimates appear to be more realistic under the given conditions.