1H high-field nuclear magnetic resonance (NMR) was used to analyze 216 extra virgin olive oils collected in three years (1996, 1997, and 1998) in different Italian areas in order to evaluate the potential contribution of this technique to the geographical characterization of olive oils. A statistical procedure performed on the intensity of selected NMR peaks has been proposed. Tree clustering analysis of NMR data performed without any a priori hypothesis showed the existence of reliable parameters able to group the olive oils according to the location of olive oil production. Linear discriminant analysis applied to selected NMR parameters of olive oils of the same year of production allowed the grouping of samples according to their geographical origin with only very few errors. Moreover, a satisfactory grouping is reached by combining the NMR data of olive oils from two different years (1996 and 1997). Operating on appropriate sampling, a careful analysis of data yielded the conclusion that the place of olive production could be singled out as a discriminating factor regardless of the cultivars from which the olive oils are derived.