BACKGROUND AND PURPOSE: Lay people often mention stress as one of the most important risk factors for stroke. Stress might trigger a cerebrovascular event directly or could be associated with higher levels of blood pressure or an unfavorable lifestyle. To examine these possibilities, we analyzed the association between self-reported stress frequency and intensity and risk of stroke. METHODS: Data from the second examination, 1981 to 1983, of participants in the Copenhagen City Heart Study were analyzed with Cox regression after a mean of 13 years of follow-up. A total of 5604 men and 6970 women were included, and 929 first-ever strokes occurred, of which 207 (22%) were fatal within 28 days after onset of symptoms. The stress frequency categories were never/hardly ever, monthly, weekly, or daily. The stress intensity categories were never/hardly ever, light, moderate, or high. RESULTS: Subjects with high stress intensity had almost a doubled risk of fatal stroke compared with subjects who were not stressed (relative risk [RR], 1.89; 95% CI, 1.11 to 3.21). Weekly stress was associated with an RR of 1.49 (95% CI, 1.00 to 2.23). There was no significant effect of stress in analyses of nonfatal strokes. Subjects who reported to be stressed often were more likely to have an adverse risk factor profile. CONCLUSIONS: Self-reported high stress intensity and weekly stress were associated with a higher risk of fatal stroke compared with no stress. However, there were no significant trends, and the present data do not provide strong evidence that self-reported stress is an independent risk factor for stroke.