AIM: The aim of this study was to assess the association between weight change and mortality due to cardiovascular diseases (CVDs) in a Japanese population. METHODS: We used the data of a population-based prospective cohort study that was conducted from 1988 to 1990 in 45 areas throughout Japan. Among a total of 69,681 men and women aged 40-79 with no history of CVD or cancer at baseline, the association between weight change from 20 years of age to baseline and CVD-related mortality was evaluated. RESULTS: During a median follow-up period of 19.1 years, we observed 4,274 deaths from total CVD. After adjusting for age, sex, and other potential confounding factors, compared with participants with a weight change of ＜2.5 kg (stable weight), participants with a greater weight change (either loss or gain) had an increased risk of mortality from total CVD (U-shaped association). The hazard ratios for the total CVD risk in participants with a weight loss and a weight gain of ≥ 12.5 kg were 1.50 (95% confidence interval [CI], 1.30-1.72) and 1.21 (95% CI, 1.07-1.36), respectively. The associations between weight change and risk of mortality from ischemic heart disease or stroke showed similar trends. The risk of intracerebral hemorrhage was associated with weight loss only. Weight change was not associated with mortality from subarachnoid hemorrhage. CONCLUSIONS: Weight loss or gain could be a risk factor for mortality from total or ischemic CVD, while weight loss could be a risk factor for intracerebral hemorrhage.