The lip biopsy is essential for the diagnosis of primary Sjogren's syndrome (SS) but an invasive method can cause some disadvantages. The purpose of this study is to apply Raman spectroscopy to detect the pathological minor salivary glands in primary SS and establish the diagnostic model of Raman spectra of the primary SS samples. Raman spectra from the primary samples and control samples were obtained by Raman microscope and were compared to find the differences. The principal component analysis (PCA) and discrimination function analysis (DFA) were employed to analyze the spectra and establish the diagnostic model. The differences of Raman spectra demonstrated the biochemical molecular alterations between the different samples. Compared with the control samples, the content of proteins, nucleic acids, and keratin increased in the primary SS samples but the content of lipids decreased. PCA and DFA displayed a powerful role in the classification of the Raman spectra. The sensitivity and specificity of the diagnostic model reached above 91 and 92%, respectively. The total accuracy is 92.4%. Raman spectroscopy combined with PCA-DFA algorithm will provide an effective and accurate technology for the diagnosis of the pathological minor salivary glands in primary SS, which may replace the lip biopsy in the future.