The present study was aimed to investigate the potential of multispectral images coupled with chemometric tools (PLSDA and PLS-R) for: (1.) discriminating different French blue veined cheeses belonging to four brand products (Fourme d'Ambert, Fourme de Montbrison, Bleu d'Auvergne, and Bleu des Causses) and (2.) predicting some of physicochemical (pH, ash, dry matter, total nitrogen, water soluble nitrogen, Ca 2+ , Na + , Cl â , and P) and rheological properties (softening and dropping points). The results obtained showed that multispectral imaging system applied to anisotropic blue cheeses succeeded to: (1.) discriminate cheeses based on their blue veins features in spite of the visual similarity of their structure and appearance with percentage of correct classification varying between 30 and 100%; and (2.) predict selected parameters (i.e., Ca 2+ , Cl â , WSN, dropping, and softening points) since R 2 cv â„ 0.62 and RPD â„ 1.62 were obtained. Moreover, the predictive results suggested that the image texture of cheese was strongly related to its physicochemical composition and rheological characteristics (softening and dropping points).