Objective The study aims to examine the association between salt preference and mortality from stroke and coronary heart disease (CHD). Methods Between 1988 and 1990, 35515 men and 49275 women aged 40–79 years completed a self-administered questionnaire in the Japan Collaborative Cohort Study for Evaluation of Cancer Risk sponsored by Monbusho. During a median duration of 16.4 years, 1970 stroke and 922 CHD deaths were observed. Salt preference was divided into three groups: low, moderate and high. Results Mortality rates per 1000 person-year from stroke were 2.0 for men, 1.3 for women and 1.6 for total subjects. The respective mortality from CHD was 1.1, 0.5 and 0.8, and that from total cardiovascular disease was 4.6, 2.9 and 3.6. Salt preference was positively associated with mortality from stroke for both sexes. The multivariable hazard ratios of stroke mortality for high versus low salt preference were 1.21(0.99–1.49) for men, 1.22(1.00–1.49) for women and 1.23(1.06–1.41) for total subjects. That positive association was primarily observed among male heavy drinkers (≥ 46.0 g ethanol/day). Salt preference tended to be inversely associated with mortality from CHD. Conclusion Salt preference was associated with increased mortality from stroke for both sexes, particularly for male heavy drinkers.