Abstract: Several mental disorders have consistently been found high likelihood to co-occur with alcohol use, but the nature of such association is still an unresolved issue. To facilitate public health and the intervention procedure for clinical treatment, the main question we are trying to answer is whether there is a causal effect of alcohol use on mental disease. We applied two-sample Mendelian randomization（MR） analysis, using the newest publicly available summary statistics from genome-wide association studies（GWAS） on five major mental disorders, including major depressive disorders（MDD）, bipolar disorders（BPD）, schizophrenia（SCZ）, autism spectrum disorder（ASD）, attention-deficit hyperactivity disorder（ADHD） and alcohol use（alcohol dependence, alcohol problems, and drinks per day）. The analyzing method was inverse-variance weighted（IVW）. MR-Egger regression and weighted median（WM） analyses were further used as sensitivity analyses. We found evidence for significant causal effects of SCZ on alcohol dependence（IVW beta =0.158 5, 95% CI, 0.058 4 to 0.258 7, P=0.001 9）, and on drinks per day（IVW beta=-0.013 14, 95% CI,-0.021 87 to-0.004 417, P=0.003 2）. We also found suggestive causal evidence for ASD（IVW beta=0.425 9, 95% CI-0.784 8 to-0.067 05, P=0.020） on alcohol dependence and ADHD（IVW beta=0.225 7, 95% CI 0.023 83 to 0.427 5, P =0.028） on alcohol dependence. Our findings strongly suggested SCZ and ADHD would increase the risk of alcohol dependence. Besides, the study also unraveled the causal relationships from SCZ to alcohol use and tentatively from ASD and ADHD to alcohol use.