Surface-enhanced Raman scattering （SERS） detection systems based on silver nanomaterials can be used to improve the weak Raman signal in bioactive substances detection. In the present work, SERS technology was used to detect joint fluid samples of 18 synovial arthritis patients and 15 normal people,and the spectral peak assignment methods, principal component analysis （PCA）-linear discriminant analysis （LDA） algorithm were applied to analyze the sample data after collecting SERS spectral data. The results of this study show that higher concentrations of polysaccharide （477 cm -1）, DNA （722 cm -1）, δ（CH2） （1439 cm -1）, guanine （N3） （1576 cm -1）, and amide I bands （1676 cm -1） in arthritis patients compared to normal people, while the glycogen （490 cm -1）, phosphatidylinositol （596 cm -1）, protein tyrosine （640 cm -1）, glucose （1071 cm -1）, and protein glutamine absorptions （1645 cm -1） were comparatively less than normal people. The specificity and sensitivity of disease diagnosis by PCA-LDA algorithm were 83.3% and 80%, respectively. Consequently, the results of this research demonstrate that SERS spectroscopy based on silver nanomaterials has considerable feasibility and potential value in diagnosis and analysis of synovial arthritis disease.