Abstract: A fast and non-destructive classification method for classification of paper express document bags was established by differential Raman spectroscopy and chemometrics, utilizing the differences in paint fillers on paper express document bags. The 63 samples were cut into small pieces with the size of 0.5 cm×0.5 cm, and their Raman spectra were collected using a differential Raman spectrometer. The samples were preliminarily classified based on the differential Raman characteristic peaks of different fillers. After Z-score standardization on the original spectral data, K-means algorithm was used to subdivide the preliminary classification results. Fischer discriminant analysis was used to verify the subdivision results, and leave-one-out cross validation method was used to test the results. It was shown that the 63 samples could be divided into 4 categories based on the differences in characteristic peaks of differential Raman spectra, among which 48 samples were divided into the class I of samples with main filter of calcium carbonate. The class I of samples was divided into 3 groups by K-means algorithm, and the classification accuracy obtained by Fischer discriminant analysis reached to 93.6%, with discriminant accuracy obtained by leave-one-out cross validation method of 87.2%, indicating that the proposed method could be used for fast and accurate identification of paper express document bags.