A SVAT calibration methodology based on the optimization of surface brightness temperature has been developed (Coudert et al., 2006a; 2006b, 2007) and validated at field scale in the framework of the Alpilles-ReSeDA ([1]; http://www.avignon.inra.fr/reseda/base/) experiment. This methodology has been extended at regional scale and applied to MSG/Meteosat8 thermal infrared data. MSG/SEVIRI instrument provides the Land Surface Temperature (LST) every 15mn with a spatial resolution of about 3kmx5km over France. A model spatialization is proposed, based on a Geographic Information System in order to estimate the surface temperature at the MSG pixel scale from local SVAT simulation. The estimation of the surface water and energy fluxes and of the soil water content at a finer resolution are provided from the calibration methodology applied at the MSG pixel scale to the the spatialized model. These methods have been applied in the framework of the CITRAM experimental program in order to monitor soil water content and irrigation in agricultural zones. A comparison between methodologies is presented. The validation is done with local soil moisture measurements acquired in different instrumented fields over the region and with high spatial resolution surface temperature estimation from ASTER/TERRA satellite data