Background: Obesity particularly during childhood is considered a global public health crisis and has been linked with later life health consequences including mental health. However, there is lack of causal understanding if childhood adiposity has a direct effect on mental health or has an indirect effect after accounting for adulthood body size. Objective: To investigate the total and direct effect of childhood adiposity on later life anxiety and depression. Method: Two-sample Mendelian randomization (MR) was performed to estimate the total effect and direct effect (accounting for adulthood body size) of childhood body size on anxiety and depression. We used summary statistics from a genome-wide association study (GWAS) of UK Biobank (n=453,169) and large-scale consortia of anxiety (Million Veteran Program) and depression (Psychiatric Genomics Consortium) (n=175,163 and n=173,005, respectively). Result: Univariable MR did not indicate genetically predicted effects of childhood body size with later life anxiety (beta=-0.05, 95% CI=-0.13, 0.02), and depression (OR=1.06, 95% CI=0.94, 1.20). However, using multivariable MR, we observed that the higher body size in childhood reduced the risk of later life anxiety (beta=-0.19, 95% CI=-0.29, -0.08) and depression (OR=0.83, 95% CI=0.71, 0.97). Both univariable and multivariable MR indicated that higher body size in adulthood increased the risk of later life anxiety and depression. Conclusion: Our findings suggest that the higher body size in childhood has a protective effect on later life anxiety and depression, if obesity is not present into adulthood. Higher body size in adulthood was a risk factor for later life anxiety and depression.