〔Abstract〕 Objective To investigate the relationship between the expression of p-Stat3 and Survivin in gastric cancer and the clinical characteristics and prognosis of patients, and to establish the prediction model of postoperative survival of patients combined with alanine aminotransferase(ALT) and aspartate aminotransferase(AST). Methods Data of 133 patients undergoing gastric tumor resection were collected and followed up. The expression of p-STAT3 and Survivin in gastric cancer and precancerous tissues were detected by immunohistochemical method. Kaplan-Meier method was used to draw the survival curve. Logistic regression model combined with receiver operating characteristic curve(ROC) curve was used to describe the predictive value of multi-index combined detection on postoperative survival status of patients. The prediction model of the histogram was established using Survival and RMS packages in R Studio software. Results The expressions of p-Stat3 and Survivin in gastric cancer tissues were higher than those in precancerous tissues. The expression of p-Stat3 in gastric cancer tissues was positively correlated with the depth of invasion and TNM stage, while the expression of Survivin in gastric cancer tissues was positively correlated with the degree of differentiation and the depth of invasion. The higher the expression levels of p-Stat3 and Survivin, the worse the prognosis. The lower the ALT level, the worse the prognosis. Survivin, ALT and AST were the optimal combination for predicting postoperative survival in patients with gastric cancer. Conclusion The expression of p-Stat3 and Survivin plays an auxiliary role in the diagnosis of gastric cancer. P-Stat3, Survivin and ALT are correlated with the prognosis of patients with gastric cancer and have predictive value. Survivin, ALT and AST combined model has a big value to predict the postoperative survival prognosis of gastric cancer patients, which can provide a reference for clinical practice.