BACKGROUND Gas chromatography–olfactometry (GC–O) is the most frequently used method to estimate the sensory contribution of single odorant, but disregards the interactions between volatiles. In order to select the key volatiles responsible for the aroma attributes of Congou black tea (Camellia sinensis), instrumental, sensory and multivariate statistical approaches were applied. RESULTS Using sensory analysis, nine panellists developed eight descriptors: floral, sweet, fruity, green, roasted, oil, spicy, and off‐odour. Linalool, (E)‐furan linalool oxide, (Z)‐pyran linalool oxide, methyl salicylate, β‐myrcene, and phenylethyl alcohol, which were identified from the most representative samples by the GC–O procedure, were the essential aroma‐active compounds in the formation of basic Congou black tea aroma. In addition, 136 volatiles were identified by gas chromatography–mass spectrometry (GC–MS), among which 55 compounds were determined as the key factors for six sensory attributes by partial least‐square regression (PLSR) with variable importance of projection scores. CONCLUSION Our results demonstrated that headspace solid‐phase microextraction/GC–MS/GC–O was a fast approach for isolation and quantification aroma‐active compounds. The PLSR method was also considered to be a useful tool in selecting important variables for sensory attributes. These two strategies, which allowed us to comprehensively evaluate the sensorial contribution of a single volatile from different perspectives, can be applied to related products for comprehensive quality control. © 2018 Society of Chemical Industry