Objectives : To investigate the association between BMI and Mortality. Methods : This study was based on the analysis and assembly of the 'Kangwha Cohort Study', previously conducted by the Department of Preventive Medicine, Yonsei University. A total of 2,696 males and 3,595 females were followed for almost ten years and ten months from March 1985 to January 1996, a total of whom 2,420 died during this period. The Cox's proportional hazards regression model was used to analyze this data. Results : We found a U-shaped relationship between BMI and mortality among the aged men in the Kangwha cohort. The hazard ratio of dying was adjusted for age, marital status, occupation, self cognitive health level, chronic disease, smoking, and alcohol frequency, then sorted by body mass index into the following groups; less than 10.5, 18.5 to less than 21.0, 21.0 to less than 23.5, 23.5 to less than 26.0 and greater than or equal to 26. The corresponding ratios for men were 1.81(1.50-2.19, 95%CI), 1.31(1.14-1.51, 95%CI), 1.0(referent), 1.05(0.87-1.26, 95%CI) and 1.39(1.09-1.76, 95%CI), respectively. And for women, 1.46(1.19-1.78), 1.12(0.95-1.31, 95%CI), 1.0(referent), 1.00(0.84-1.20, 95%CI) and 1.09(0.89-1.34, 95%CI), respectively. Conclusions : The risk of death among aged men in Kangwha increased in the under and overweight groups. The relationship between BMl and mortality has been well studied in Western populations, but little is known about the association between BMI and mortality in our country. So, on the basis of this study, it is apparent that more studies of the relationship between BMI and mortality will be needed for future work.