Background and purpose: Previous observational studies have reported that patients with migraine have an increased risk of stroke. We explored whether migraine has a causal effect on stroke using a two‐sample Mendelian randomization approach. Methods: Genetic instruments were selected from large genome‐wide association studies of migraine and stroke. A two‐sample Mendelian randomization analysis was performed, along with sensitivity analysis. We used migraine subtypes (any migraine, migraine with aura, migraine without aura) as risk factors and stroke, ischemic stroke, and hemorrhagic stroke as outcomes for this analysis. Ischemic stroke subtypes were also included to explore the underlying pathogenesis linking migraine to stroke. Results: Migraine did not show any association with stroke (odds ratio [OR], 0.95; 95% confidence interval [CI], 0.87–1.03), ischemic stroke (OR, 0.93; 95% CI, 0.85–1.02), or hemorrhagic stroke (OR, 1.26; 95% CI, 0.84–1.91), suggesting that the observed association may not be causal. Neither migraine with aura nor without aura showed causal relationship with outcomes. The sensitivity analysis supported the results of the primary analysis. Regarding ischemic stroke subtypes, migraine seemed to have a negative association with large‐artery atherosclerosis (OR, 0.81; 95% CI, 0.68–0.95), whereas associations with small‐vessel occlusion or cardioembolism were not evident. Conclusions: Contrary to previous observational studies, we were unable to find any causal relationship between migraine and stroke. However, the suggested negative association of migraine in large‐artery atherosclerosis warrants further research. [ABSTRACT FROM AUTHOR]