Abstract: 21 jet fuel samples collected from different refineries around different regions in China were took as the research objects. Comprehensive two-dimensional gas chromatography-mass spectrometry was used for qualitative and semi-quantitative analysis. A comprehensive two-dimensional gas chromatography coupled with single quadrupole mass spectrometry (GC×GC-qMS) equipped with polar-nonpolar column system can significantly increase the peak capacity, which had obvious advantages in analyzing jet fuel with complex components. Based on percentage response (the percentage of each peak volume to all peak volumes), the method of F-ratio (the ratio of the average square error between the dataset to the average square error within the dataset) in analysis of variance (ANOVA) was used to select feature components of jet fuel samples. The results showed that monocyclic aromatic hydrocarbons, indane series, tetralin series with number of carbon atoms in the range of C₁₀-C₁₂ varied greatly between different groups of jet fuel samples. Different methods for determining the threshold of F-ratio were compared, and the Z-score method was finally used to select experimental data, in which case a total of 16 feature components with Z-score were greater than 3 based on the percentage response F-ratio of the samples. Ultimately, combined with the extracted ion current chromatogram (EIC) with m/z 117-118 or m/z 131-132, the traditional one-dimensional gas chromatography-mass spectrometry (GC-MS) and GC×GC-qMS were compared with effectiveness of separation and analysis of the feature components. The detected peaks were obviously more than the former, and the extracted ions in this mass-to-charge ratio range had a certain distinguishing effect on jet fuel samples. This paper provided a method for comprehensive analysis of jet fuel components, and also provided a reference for selecting useful data while preserving the features of the largest difference between groups, reducing the large amount of experimental data generated in GC×GC-qMS.