Objective: To assess the casual effect of childhood obesity on adulthood coronary artery disease (CAD) using Mendelian randomization (MR) method. Methods: Data on BMI of children aged 2-10 years in 2015 were downloaded from Early Growth Genetics Consortium and Genetic Investigation of Anthropometric Traits Consortium. Twenty-seven genetic variants related to children’s BMI were selected as instrumental variables (IVs), and the associations between IVs and CAD were extracted from a Meta-analysis of the genome-wide association study of CAD cases published in UK Biobank 2015. We used MR-Egger regression to test whether there was the pleiotropy of the selected SNPs. In the present MR methods, we conducted MR analyses by using mode-based estimate method as primary method for summary-level of associations to estimate the causal association between childhood obesity and CAD. Results: The intercept term estimated for CAD from MR-Egger method suggested that the selected SNPs don’t exert pleiotropy with a 95%CI including the null (-0.008-0.018). In addition, we found evidence that support the effect of childhood obesity on CAD risk: a 1 s increase in children BMI (kg/m²), and the risk of suffering from CAD in adulthood increased by an average of 37% (OR=1.37, 95%CI: 1.09-1.72). Conclusion This study provides a causal association between childhood obesity and CAD risk.