The present paper deals with chemometric interpretation of soil analysis data collected from 31 sampling sites in the region of Kavala and Drama, Northern Greece. The determination of 16 different chemical and physicochemical characteristics is principally needed for prognosis of the land treatment and fertilizing. The study carried out indicates that the application of multivariate statistical approaches could reveal new and specific information about sampling sites. It has been found that they could be divided into four general patterns: pattern 1 contains dominantly inorganic and alkaline soil samples from semi-mountainous regions in close proximity to the seacoast; pattern 2 indicates the same soil sample type and regional location as pattern 1 but is far from the coastal line; pattern 3 includes samples from sites from the plains with organic and alkaline soils with close proximity to the coast; pattern 4 resembles pattern 3 as soil type but involves samples from sites far from the shore. Further, six latent factors were identified, conditionally named “structural”, “acidic”, “nutritional”, “salt”, “microcomponents” and “organic”. Finally, an apportioning procedure was carried out to find the source contributions in the measured analytical values. In this way the routine estimation of the soil quality could be improved.