Background: Heavy drinking is a risk factor for hypertension. Alcohol consumption is also associated with weight changes. Therefore, the purpose of this study is to examine the mediating effect of weight change on the incidence of hypertension depending on the level of alcohol consumption in a cohort study of Japanese men. Methods: The participants were 2,137 non-hypertensive male employees (18-61 years old, mean age 38.3) of a metal products factory in Japan. BMI and blood pressure (BP) were measured during a health examination in 2009. BP was measured twice after resting, and the means of BP values were analyzed. Types of alcohol, amount and frequency of drinking were evaluated using a self-administered questionnaire. Follow-up body weight and BP were measured during a health examination in 2014. The incidence of hypertension was defined as systolic BP 130 or higher, diastolic BP 80 or higher, or oral antihypertensive medication. We investigated the mediating effect of BMI on the relationship between alcohol consumption and blood pressure. Associations were adjusted for baseline age, BMI, and systolic BP. Results: The mean values (standard deviations, SD) of baseline BMI (kg/m 2 ), systolic and diastolic BP (mmHg) were 22.6 (2.7), 111.6 (9.2), 67.0 (7.3), respectively. During the 5-year follow-up, 429 incident case of hypertension were identified. The mean body weight change (kg) over 5 years for each alcohol consumption category (non-drinker, <20g ethanol/day, 20-39 g/day, and ≥40 g/day) was 0.8 (4.4), 0.7 (4.0), 0.0 (3.8), and -0.4 (4.4), and the incidence rates (%) of hypertension were 16.0, 17.7, 28.0, and 32.1, respectively. The multivariate-adjusted odds ratios (95% confidence intervals) for incident hypertension were 1.00 (reference), 1.22 (0.90-1.64), 1.71 (1.19-2.47), and 2.06 (1.37-3.09); ORs were significantly higher for those with 20-39g/day and ≥40g/day groups compared to non-drinkers. In the mediation analysis, the odds ratios for the direct effect of each alcohol consumption group on the development of hypertension were 1.00 (reference), 1.27 (0.94-1.73), 1.91 (1.39-3.44), 2.36 (1.56-3.58), and those for indirect effect mediated by body weight change were 1.00 (reference), 0.99 (0.95-1.04), 0.96 (0.90-1.01), and 0.93 (0.85-0.99), respectively. Conclusion: Alcohol consumption was inversely associated with body weight change and positively associated with the incidence of hypertension. Among heavy drinkers, there was a significant indirect effect mediated by body weight, and weight loss due to heavy drinking lowered the risk of developing hypertension. When evaluating the relationship between alcohol consumption and the risk of hypertension for heavy drinkers, the risk due to alcohol drinking may be underestimated through changes in body weight.