〔Abstract〕 A descriptive cohort study was conducted to analyze purine metabolism in saliva samples from 178 patients with gastritis and gastric cancer. 167 differential metabolites were identified by ultra-high performance liquid chromatography-mass spectrometry(UHPLC-QTOFMS). Xanthosine, Cyclic GMP, uric acid and 2′-Deoxyadenosine 5′-monophosphate were screened out by Studentttest and orthogonal partial least square analysis. ROC curve was used to evaluate the inter group discrimination effect of diagnostic markers. Fisher discriminant analysis was used to establish the diagnostic function model, and the sensitivity and specificity of the diagnostic model were obtained as 83.1% and 63.3% respectively by leave-one-out cross validation. The screening of purine metabolites in saliva has great potential in distinguishing early gastric cancer from gastritis and screening early gastric cancer.