BACKGROUNDThere is an association between thyroid dysfunction and cerebral infarction (CI), but the causality cannot be determined. A two‐sample two‐way Mendelian randomization (MR) study was conducted to assess the causal relationship between thyroid function and CI.METHODSWe selected single‐nucleotide polymorphisms (SNPs) associated with five phenotypes, including CI from the UK Biobank (n = 361,194), hyperthyroidism from the IEU Open GWAS database (n = 484,598), hypothyroidism from the IEU Open GWAS database (n = 473,703), normal thyroid‐stimulating hormone (TSH) (n = 271,040), and normal free thyroxine (FT4) (n = 119,120) from the Thyroidomics Consortium database. For the forward MR analysis, the exposures were hyperthyroidism, hypothyroidism, TSH, and FT4. The inverse variance weighted (IVW) method, weighted median (WM), and MR‐Egger revealed the causality with CI. For the reverse MR analysis, CI was regarded as the exposure, and four thyroid function phenotypes were the outcomes. The sensitivity and heterogeneity test was assessed using Cochran's Q test, MR‐Egger regression, and leave‐one‐out analysis.RESULTSThe MR analysis indicated that genetic susceptibility to hyperthyroidism increased the risk of CI (IVW‐OR = 1.070; 95% CI: 1.015–1.128; p = 0.003). In reverse MR, genetic susceptibility to RA is not associated with hyperthyroidism (IVW‐OR = 1.001; 95% CI: 1.000–1.001; p = 0.144). Any positive or reverse causal relationship between hypothyroidism, FT4, and TSH with CI could not be established. Sensitivity and heterogeneity test consolidated our findings.CONCLUSIONThe causality between CI and hyperthyroidism demonstrated patients with hyperthyroidism have a risk of genetic variants for CI. In the future, further studies are needed to fully explore their mechanisms of action.