Background Endoscopic screening has been adopted in South Korea for the national screening of gastric cancer (GC). This study aimed to assess the effect on overall survival of GC patients and determine the optimal endoscopic screening interval. Methods The baseline characteristics and overall survival of GC patients treated at the National Cancer Center, Korea, between 2010 and 2016 were compared between those without a history of endoscopic evaluation (group N) and those in whom the interval between the last endoscopic evaluations and diagnosis of GC was ≤ 1, 1–2, 2–3, 3–4, or > 4 years (groups 1–5, respectively). Results A total of 2362 patients met the criteria for the study (1060 in group N and 1302 in groups 1–5). More patients in groups 1–5 were diagnosed with stage I GC (83.7, 83.7, 71.8, 78.2, and 71.6%, respectively) than in group N (62.4%, P  < 0.001) and were treated endoscopically (38.8, 33.8, 24.7, 21.8, and 15.5%, respectively, vs. 13.5%; P  < 0.001). Group 2 had less-advanced tumor stages ( P  = 0.001) and was more likely to have received endoscopic treatments ( P  = 0.026) than group 3. Hazard ratios for death were significantly lower in groups 2 (0.45; 95% confidence interval CI, 0.32–0.64) and 3 (0.57; 95% CI, 0.33–0.98) than in group N; the decrease was not significant in group 4 (0.49, 95% CI, 0.20–1.20). Conclusions Endoscopic screening every 3 years may reduce the mortality of GC patients, though screenings at least every 2 years may benefit patients with less-advanced stages.