The fluorescent fingerprints of diesel oil samples were investigated in order to develop a fast and cost-effective method to facilitate the discrimination of rebated tax diesel fuel from oil that is illegally processed by the sorption process. In the experiment oil samples were spiked with a fiscal marker (Solvent Yellow 124) and a dye (Solvent Red 19) and then these were removed using a simulated sorption process. The excitation-emission fluorescence fingerprints were recorded for each sample. Discriminant models were constructed on the basis of fluorescence spectra to distinguish oil samples with respect to four possible discrimination schemes (corresponding to the concentrations of chemical additives). Using discriminant partial least squares models, in all of the discrimination cases that were considered, 100% of the samples from the model set were discriminated correctly. Prediction results for the test set samples were encouraging and varied between 77% and 100% of correctly discriminated samples. (C) 2013 Elsevier Ltd. All rights reserved.