BACKGROUND: The causal relationships between gut microbiota, blood metabolites, and stroke and its subtypes remain unclear. This study aims to uncover the causal associations using Mendelian randomization. METHODS: We initially identify Single-Nucleotide Polymorphisms (SNPs) correlated with gut microbiota and blood metabolites as instrumental variables (IVs) from the summary statistics in Genome-Wide Association Study (GWAS) to evaluate their potential causal associations with stroke and its subtypes. We proceed with a two-step Mendelian randomization analysis aiming to determine whether blood metabolites mediate the relationships between gut microbiota and stroke or its subtypes. RESULTS: We identified the genetic predictions of 12, 11, and 10 particular gut microbiota were associated with stroke, ischemic stroke, and intracerebral hemorrhage respectively. Inverse variance weighted (IVW) analysis disclosed Alistipes (OR [95%CI]: 1.11[1.00,1.23]), Streptococcus (OR [95%CI]: 1.17[1.05,1.30]), and Porphyromonadaceae (OR [95%CI]: 2.41[1.09,5.31]) as the primary causal effects on stroke, ischemic stroke, and ICH, respectively. We determined that 8, 11, and 1 blood metabolites were causally related to stroke, ischemic stroke, and ICH, respectively. Among these metabolites, Citrate (OR [95%CI]: 2.39[1.32,4.34]) and Beta-hydroxyisovalerate (OR [95%CI]: 2.54[1.62,3.97]) had the foremost causal effect on stroke and ischemic stroke, respectively, whereas Glutaroyl carnitine evidenced a causal effect on ICH. Furthermore, our study revealed that Tetradecanedioate marginally mediated the causal effects of Paraprevotella on stroke and ischemic stroke. CONCLUSIONS: This study established a causal link between gut microbiota, plasma metabolites, and stroke. It revealed a marginal pathway, shedding new light on the intricate interactions among gut microbes, blood metabolites, stroke, and their underlying mechanisms.