Background <p>The association of fish/seafood consumption with stroke remains inconclusive.</p> Methods <p>This cohort included 53,916 participants aged 30–79 years from the China Kadoorie Biobank prospective study in Tongxiang. Incident stroke was obtained periodically through linkage with stroke surveillance system, death registries, and national health insurance system. Cox regressions were used to estimate the association between the frequency of fish/seafood consumption and incident stroke or its subtypes.</p> Results <p>48.2% of participants reported consuming fish/seafood at least once per week. During 643,100 person-years (median 12.4 years) of follow-up, 2994 total strokes, 2213 ischemic strokes (IS), and 772 hemorrhagic strokes (HS) were identified. Among all participants, fish/seafood consumption was not associated with the risk of total stroke or its subtypes. Among men, in comparison with non-consumers, the adjusted HRs (95%CI) for total stroke risk were 0.90 (0.74–1.11) for monthly, 0.83 (0.68–1.03) for 1–3 days/week, and 0.64 (0.45–0.91) for ≥ 4 days/week consumption (<i>p</i>-trend = 0.009). A similar protective pattern was observed for IS (<i>p</i>-trend = 0.024), whereas no association was observed with HS. Among women, no significant association was observed between fish/seafood consumption and total stroke or its subtypes.</p> Conclusions <p>Fish/seafood consumption was found to be unrelated to HS. Higher fish/seafood consumption was associated with lower risks of total stroke and IS in men but not in women. Future studies need to validate these subtype- and sex-specific associations.</p>