peer reviewed ; Data acquired from field campaign and hyperspectral airborne sensors were processed to determine the surface soil organic matter of an agricultural area located in Southern Belgium. The method adopted was based on a forward stepwise multiple linear regression analysis linking soil organic matter and hyperspectral data from the CASI-2 (Compact Spectrographic Imager-2) airborne sensor working in the visible and near infrared domain. The results were validated successfully from an independent set of sampling points. However, disturbing factors effects are shown on the relationship between soil organic matter and spectral reflectance. It is concluded that the hyperspectral remote sensing approach is promising for soil organic matter prediction but it will require more study to better take account on the disturbing factors affecting the relationship.