This work presents a statistical calibration of a semi-empirical combined optical-passive microwave remote sensing method to estimate surface soil moisture at regional scale. The proposed method uses satellite data provided from different sensors such as: bi-polarized brightness temperature at L-band given by the Soil Moisture Ocean Satellite (SMOS) level 2 product (SMUDP2), the Normalized Difference Vegetation Index (NDVI) and Land surface Temperature (LST) derived from MODIS (MOD13Q1 and MOD 11A1 respectively), the LST obtained by GOES GEOLAND2 products, the surface soil moisture (Volumetric soil water layer 0 – 7 cm) and the skin temperature provided by the ECWMF (ERA-interim). The NDVI was interpolated at temporal (daily) and spatial (from 250 m to 10 km) resolutions in order to assess the influences of this vegetation indicator on the semi-empirical approach. Results shown the use of vegetation index such as the NDVI improve the calibration of the semi-empirical approach. The improvements of the coefficient of determination (r2) increase between 2 and 20% at using NDVI in comparison to setting NDVI equal to zero. The use of LST in the calibration of the semi-empirical approach can be considered as a possible indicator of the effective temperature. The three LST data sources have not shown high differences in the calibration approach (p-value > 0.05) in relation to the use of effective temperature. A partial evaluation of soil moisture estimates were performed without statistical significance results. Finally, this work contributes to the use of synergic optical-passive microwave approach which can improve the soil moisture estimation based on remote sensing data at regional scale.< Réduire