Abstract:A method to detect 13 polyphenols in wine and research models of wine identification was established by high pressure liquid chromatograph-quadrupole/electrostatic field orbitrap high resolution mass spectrometry and multivariate statistical analysis technology.The samples were extracted with ethyl acetate after acidizing.The qualitative confirmation was carried out by full scan and target MS2 spectra after separated on a ACQUITY UPLC BEH C18column.An external standard method was used for quantitative analysis.The obtained data were analyzed by principal component analysis and partial least-squares discriminant analysis based on information of production area，composition and country.The results showed that 2，5-dihydroxybenzoic acid and vanillic acid were important indicators to distinguish red wine and white wine，while trans-resveratrol and 2，5-dihydroxybenzoic acid were important indicators of different production area.The accuracy of the corresponding models was 96.2%and 55.3%，respectively.It could preliminarily realize the identification of wine quality and producing area.