Although numerous studies have been done on gastric cancer and alcohol consumption, results from these studies are inconsistent. We conducted a population-based, prospective cohort study to establish a relationship between alcohol and gastric cancer according to gender. The cohort consisted of elderly (>64 years of age) subjects at the baseline of 1993-1998. Baseline information was surveyed using a self-administered questionnaire. Gastric cancers were identified by the National Cancer Registry. Cox proportional hazards models were used to calculate relative risks with 95% confidence intervals. A follow up of 116,997.1 person-years of the 13,396 subjects revealed 151 newly diagnosed gastric cancers (80 men and 71 women). The risk of gastric cancer was higher by a factor of three among female former drinkers (adjusted relative risk 2.85 [95% confidence interval 1.11-7.32]) compared to current and nondrinkers. Female former drinkers showed greater alcohol consumption than current drinkers (36.5g/week vs. 16.4g/week; P<.0001) and a longer duration of alcohol consumption than did current drinkers (24.5 years vs. 18.46 years; P<.0001). Female subjects with more than 110g of weekly alcohol consumption had an increased risk of developing gastric cancer (adjusted relative risk 2.23 [95% confidence interval 0.79-6.29]), although the result was statistically insignificant. No relationship was observed for male subjects. The relationship between alcohol and gastric cancer differs according to gender. Alcohol consumption may increase the risk of gastric cancer in women, and the risk elevation may persist for several years after drinking ceases.