Organoleptic properties and particularly aroma of dark chocolate depend on cocoa origin, cocoa variety and fabrication process. A sensory analysis performed on 206 dark chocolates produced from cocoa beans of different varieties and origins but with the same fabrication process classified them into four sensory categories. The objectives of this work were i) to assess whether the aromatic composition of the chocolates determine the sensory category they belong to and ii) to identify the key odorants that allow the dark chocolates discrimination. For this, volatile organic compounds (VOCs) emitted from the samples were analyzed by dynamic headspace coupled to a direct-injection mass spectrometry method using a Proton Transfer Reaction – Time of Flight – Mass Spectrometer (PTR-ToF-MS). The analyses of 1 g of chocolate mixed with 1 mL of artificial saliva in 20 mL vials were performed in triplicate under stirring at 36.2°C after 2 hours equilibration time. The average areas under the curves obtained for 314 significant ions present in the mass spectra during 2 mn release time were used to perform supervised and unsupervised multivariate data analyses. We showed this headspace PTR-MS analyses allowed retrieving the classification of the 206 samples into the four sensory categories previously determined. In a second time, to determine the key aroma of each sensory category, a Gas Chromatography-Olfactometry (GC-O) study of extracts representative of each subset of chocolates was undertaken. Twelve samples (3 samples per sensory category) among the 206 were selected. The aroma fractions were isolated by solvent assisted flavour evaporation (SAFE) from 30 g of chocolate diluted in 100 mL of water. The distillates were extracted with methylene chloride and the extracts were submitted to GC-O using the detection frequency method. The odour events identified by a panel of 12 assessors were grouped into olfactive areas (OAs). On average 50 OAs were found in the extracts and th