This study aimed to explore the application of surface-enhanced Raman scattering (SERS) in the rapid diagnosis of gastric cancer. The SERS spectra of 68 serum samples from gastric cancer patients and healthy volunteers were acquired. The characteristic ratio method (CRM) and principal component analysis (PCA) were used to differentiate gastric cancer serum from normal serum. Compared with healthy volunteers, the serum SERS intensity of gastric cancer patients was relatively high at 722cm-1, while it was relatively low at 588, 644, 861, 1008, 1235, 1397, 1445 and 1586cm-1. These results indicated that the relative content of nucleic acids in the serum of gastric cancer patients rises while the relative content of amino acids and carbohydrates decreases. In PCA, the sensitivity and specificity of discriminating gastric cancer were 94.1% and 94.1%, respectively, with the accuracy of 94.1%. Based on the intensity ratios of four characteristic peaks at 722, 861, 1008 and 1397cm-1, CRM presented the diagnostic sensitivity and specificity of 100% and 97.4%, respectively, and the accuracy of 98.5%. Therefore, the three peak intensity ratios of I722/I861, I722/I1008 and I722/I1397 can be considered as biological fingerprint information for gastric cancer diagnosis and can rapidly and directly reflect the physiological and pathological changes associated with gastric cancer development. This study provides an important basis and standards for the early diagnosis of gastric cancer.