Sensory characteristics of Virgin Olive Oil (VOO) strongly influence consumer preferences. These characteristics are also affected by the varietal origin of the olives used for producing the oils, and consequently a reliable approach is strongly required to authenticate the varietal origin of VOO. However, an approach is still missing to differentiate VOOs of the most worldwide spread cultivars, thus protecting consumers and producers from frauds. The aim of this research was to propose chemometric models for discriminating monocultivar virgin olive oils of nine widespread cultivars. 72 volatile compounds of 320 monocultivar samples from 4 crop seasons and several geographic areas were quantitated by a validated HS-SPME-GC-MS method. The VOCs able to discriminate among cultivars were selected through ANOVA: most of them belonged to the lipoxygenase pathway. A PLS-DA model was successfully validated, with percentages of correct classification up to 95.9% after external validation. Three clusters of cultivars were then defined using Agglomerative Hierarchical Clustering and Correlation Analysis. LDA and PLS-DA models correctly classified 94% of samples in the three clusters. The results indicated that the volatile profile of samples of diverse cultivars are not completely independent from each other, and that specific groups of cultivars showed strong similarities not given by the geographic origin. Overall, this study will contribute to protect from frauds and to valorize monocultivar extra virgin olive oils.