Southeastern Mexico, particularly Tabasco's flatlands which experienced a severe flood in 2007, was used as a case study for testing a methodology for the estimation of direct damage looses on agricultural crops by flooding. We proposed an accurate delineation of agricultural lands of multispectral images (SPOT-5) which consist on ensemble classifiers trough a majority voting, that combine spatial and spectral information. Finally in order to evaluate the impact of floodwater, a radar data (RADARSAT-1), were used for both, delineating the flood extent and estimating water depth. These layers were overlaid on the agricultural crop classification layer, and crop yield damage was estimated using a depth damage function. The results of this research quantified and evaluated the overall economic loss (tangible damage) from the impact of floodwater on agricultural crops.