Abstract: The purpose of this study was to develop a rapid non-destructive method for the detection of polysaccharides in dry red wine by attenuating total reflection-Fourier transform infrared spectroscopy (ATR-FTIR) combined with chemometric methods.In the experiment, 100 dry red wines from Xinjiang production area were used as test materials, wine polysaccharides were extracted by alcohol precipitation method, polysaccharide powder was obtained by vacuum freeze-drying, and then the polysaccharides were decomposed into monosaccharides with trifluoroacetic acid, and the monosaccharides were quantitatively analyzed by high performance liquid chromatography (HPLC-PDA), and the content of different types of polysaccharides in the wine samples was calculated according to the monosaccharide concentration of their respective characteristic structures according to a certain molar ratio, including total soluble polysaccharides (TSP), mannosan protein (MP), arabinose-galactose-rich polysaccharides (PRAG), rhamnogalacturonic acid glycan type II (RG-II), homogalacturonized glycan (HG), and dextran (GL).The mid-infrared spectral information of wine was collected by ATR-FTIR, and the spectral preprocessing was carried out by standard normal transform (SNV) and multivariate scattering correction (MSC) methods, followed by the competitive adaptive reweighting algorithm (CARS) for band screening, and finally combined with two modeling methods, partial least squares regression (PLSR) and backpropagation neural network (BPNN), modeling, prediction, and evaluation indicators were fitted with spectral characteristic information in the 1900～900 cm⁻¹ band to determine the content of several polysaccharide substances measured by HPLC-PDA. The results indicate that the polysaccharide content of different categories varied greatly between the test liquor samples, with TSP content of (859.41±293.65) mg/L, MP (208.08±78.42) mg/L, PRAG (418.30±140.00) mg/L, RG-II (113.17±55.11) mg/L, GL (95.46±62.10) mg/L, and HG (24.41±55.86) mg/L. The characteristic information screened from 1900～900 cm⁻¹ was used to fit the content of several polysaccharides in the test wine samples, modeling by using the linear and nonlinear correction methods, the results showed that the ATR-FTIR model has good predictive power for the content of several classes of polysaccharides in wine.among, The PLSR model showed a better prediction performance than the BPNN, The coefficient of determination (Rc²) of the PLSR model between the characteristic bands and content of polysaccharides (TSP, MP, PRAG, RG-II, and GL) was 0.98,0.96,0.92,0.99,0.98, respectively, The coefficient of determination (Rp²) is 0.85, 0.92, 0.83, 0.83, 0.84, respectively, The relative analysis error (RPDc) in the training set is 6.50, 5.31, 3.62, 9.10, and 7.86, respectively, The relative analysis error (RPD_P) of the prediction set was 2.68, 3.99, 2.44, 2.52, and 2.37, respectively. The ATR-FTIR method developed in this study for the detection of polysaccharides in dry red wine can accurately predict the content of polysaccharides in TSP, MP, PRAG, RG-II and GL by using the spectral information of the polysaccharide characteristic band of 1900 ～900 cm⁻¹, which has the application potential of rapid and nondestructive detection of polysaccharides in dry red wine.