Upper gastrointestinal bleeding is a major adverse event of non-steroidal anti-inflammatory drugs (NSAIDs), and co-administration of proton pump inhibitors and H2 receptor antagonists has been established as a means of preventing such an effect. However, the incidence of bleeding associated with NSAID-induced ulcers under conditions where such strong anti-acid agents are used for prevention has yet to be clarified. We aimed to determine the annual incidence of serious upper gastrointestinal ulcer bleeding among Japanese patients in whom NSAIDs were used in our hospital. Before commencing the study, we recommended to all the physicians in our hospital the best method for caring for NSAID users, focusing on the concomitant use of proton pump inhibitors or H2 receptor antagonists. We conducted a cohort study involving 17,270 patients for whom NSAIDs had been newly prescribed. Bleeding from gastric ulcers was observed in 8 of the 17,270 patients using NSAIDs (0.05%). The pooled incidence rate for bleeding was calculated as 2.65 (95% confidence interval, 2.56-2.74) and 1.29 (1.27-1.31) per 1,000 patient years for low-dose aspirin and non-aspirin NSAID users, respectively. None of the bleeding ulcer patients required blood transfusion or were in serious condition. In conclusion, gastric ulcer bleeding occurred in low-dose aspirin or non-aspirin NSAID users, but its incidence was low and outcomes were not serious when adequate preventive measures were taken.