Abstract: In order to objectively identify the chemical fingerprints of tobacco extracts from different growing areas, solid-phase supported liquid-liquid extraction(SLE) and gas chromatography-mass spectrometry(GC-MS)were used to analyze the volatile and semi-volatile components in the tobacco extracts from Yunnan, Zimbabwe and other areas. Combined with chemometrics, the differences of target components in the tobacco extract samples were investigated, and the characteristic components that could be used to characterize tobacco extracts from these different areas were screened out based on the analytical techniques. The results showed that: 1)In total of 125 components were identified, including 19 alcohols, 26 esters, 4 aldehydes, 22 ketones, 7 acids, 19 alkanes, 3 alkenes, 5 phenols, 8 amines, 2 ethers and 9 other compounds, in addition to nicotine which had the highest mass fraction. 2)Principal component and cluster heat map analysis could not effectively distinguish three groups of the tobacco extract samples. When the orthogonal partial least squares discriminant analysis(OPLS-DA)prediction model was applied, and effective distinguishment of the different tobacco extract samples was achieved with the identity of the three tobacco origins. The accuracies of the model training and prediction were 100%. 3)Sixteen differential compounds were identified by variable importance projection value(VIP > 1)and one-way ANOVA(P < 0.05), in which three, one and two characteristic components were mapped from the extract samples from Yunnan, Zimbabwe and the other areas, respectively. With the advantages of objectivity and high accuracy, this method could be used for quality, sample identification and origin traceability of different tobacco extracts.