Aim To study the causal relationship between coffee consumption and atherosclerosis risk. Methods Based on the public genome database IEU Open GWAS project website, relevant data were obtained, and instrumental variables closely related to exposure and outcome were screened. Mendelian randomization analysis was performed using the inverse variance weighted method, weighted median method, and MR Egger method to assess the causal relationship between coffee consumption and arteriosclerosis (including coronary atherosclerosis, peripheral atherosclerosis, cerebral atherosclerosis, and other atherosclerosis). In addition, the Egger intercept method was used to detect the level of pleiotropy, the Cochran Q test was used to assess the heterogeneity, and the leave one out method was used to perform sensitivity analysis, so as to ensure the robustness of the results. Results The study showed that the F-values of the instrumental variables included in the study were all greater than 10, and there was no weak instrumental variables bias. Coffee consumption was positively correlated with the risk of coronary atherosclerosis (OR＝1.535 5,5%CI＝1.108 4～2.127 2, P＝0.009 9), peripheral atherosclerosis (OR＝2.098 6,5%CI＝1.182 2～3.725 7, P＝0.011 4), and other atherosclerosis (OR＝1.864 7,5%CI＝1.052 9～3.302 4, P＝0.032 6), but not with the risk of cerebral atherosclerosis. Heterogeneity test Q_pval＞0.05, level pleiotropy test interval P＞0.05. Conclusions The single nucleotide polymorphisms selected in the study are all strong instrumental variables, and there is no heterogeneity or horizontal pleiotropy, indicating reliable results. The research shows that coffee consumption increases the risk of coronary atherosclerosis, peripheral atherosclerosis and other atherosclerosis, and has no significant correlation with the risk of cerebral atherosclerosis.The health management of atherosclerosis patients should reduce coffee consumption.