The biochemical composition of grape berries depends on the cultivar genome and is influenced by environmental conditions and growing practices, which vary according to origin and "terroir" (French word accounting for the factors of climate, soil, and cultural practices on grape and wine quality), The components currently measured to determine the potential quality of grapes for winemaking at harvest are sugars, acidity, pH, and total phenolics, referred to as "classic analysis". The aim of this work was to establish metabolic profiles using both conventional physicochemical analyses and H-1 NMR spectrometry of the skin and pulp of mature berry extracts in order in four appellations situated in different locations in southernwestern France (Bordeaux). Principal component analysis was applied to the physiochemical and H-1 NMR data to investigate the variability of the grape composition and to characterize groups of samples. A significant clustering of the metabolic profile of pulps or skins in relation to their terroir was observed. Physicochemical analyses were more discriminant than H-1 NMR data, but NMR spectroscopy allowed metabolic fingerprintings using identified metabolites and some still nonattributed resonances.