This study focuses on the application of the Danish national soil Visible Near Infrared Re-flectance spectroscopy (NIRs) database for predicting SOC in a field. The Conditioned Latin hypercube sam-pling (cLHS) method was used for the selection of 120 soil profiles based on DualEM21s and DEM data (ele-vation, slope, profile curvature). All the soil profile cores were taken by a 1 m long hydraulic auger with plastic liners inside. A Labspec 5100 equipped with a contact probe was used to acquire spectra at (350-2500 nm) in each 5 cm depth interval. The results show that after the removal of moisture effect using an external parameter orthogonalisation algorithm, most of the spectra collected at field moisture content can be projected in the National spectra library. Moreover, the prediction of SOC improved compared to the model based on absorbance spectra.