A sensitive and simple method was developed for the classification of wines according to variety, geographical origin, and vintage using NMR-based metabonomics. Polyphenol-rich extracts were prepared from 67 varietal wines from the principal wine-producing regions of Greece, using adsorption resin XAD-4. 1D (1)H NMR spectra obtained from the corresponding extracts were segmented, integrated, and normalized, and the data were subjected to principal component analysis. The chemometric classification of wines according to their phenolic profile allows discrimination between wines from different wineries of the same wine-producing zone and between different vintages for wines of the same variety.