Background and Purpose To test whether alcohol intake, both observational and estimated by genetic instruments, is associated with risk of ischemic and haemorrhagic stroke. Methods We used data from the Copenhagen City Heart Study 1991 to 1994 and 2001 to 2003, and the Copenhagen General Population Study 2003 to 2012 (n=78,546). As measure of alcohol exposure, self-reported consumption and genetic variation in alcohol metabolizing genes (alcohol dehydrogenase ADH1B and ADH1C) as instrumental variables were used. Stroke diagnoses were obtained from a validated hospital register. Results During follow-up 2,535 cases of ischemic and haemorrhagic stroke occurred. Low and moderate alcohol intake (1 to 20 drinks/week) was associated with reduced risk of stroke. The hazard ratios associated with drinking 1 to 6, 7 to 13, and 14 to 20 drinks/week were 0.84 (95% confidence interval [CI], 0.76 to 0.92), 0.83 (95% CI, 0.73 to 0.94), and 0.84 (95% CI, 0.73 to 0.97), respectively, compared with drinking <1 drink/day. ADH1B and ADH1C genotypes were not associated with risk of stroke. Further analysis to test the included measures revealed that increasing alcohol intake (per 1 drink/day) was positively associated with risk of alcoholic liver cirrhosis, but not associated with risk of stroke, and that increasing blood pressure (per systolic 10 mm Hg) was not associated with risk of alcoholic liver cirrhosis, but positively associated with risk of stroke. Conclusions Low and moderate self-reported alcohol intake was associated with reduced risk of stroke. The result was not supported by the result from the causal genetic analysis.