Background Several articles have shown that birth weight is associated with the risk of many types of cancers. However, the results are inconsistent and whether the relationship has a casual effect remains unknown. Objectives To estimate the association between birth weight and cancer risk by dose-response meta-analysis and two-sample Mendelian randomization analysis. Search strategy PubMed and Embase library up to March 2021. Selection criteria Prospective cohort studies and case-control studies. Data collection and analysis Two reviewers collected data and the third reviewer check the accuracy. Summary relative risks (RRs) and 95% confidence intervals (CIs) were included. Main results In our dose-response meta-analysis, six cancers from 46 studies were found to had significant associations with the birth weight. (Ovarian cancer: RR: 1.21, 95%CI: 1.01-1.44; breast cancer: RR: 1.12, 95%CI: 1.08-1.16; colorectal cancer: RR: 1.20, 95%CI: 1.01-1.43; endometrial cancer: RR: 0.85, 95%CI: 0.78-0.93; prostate cancer: RR: 1.27, 95%CI: 1.01-1.61; testicular cancer: RR: 1.21, 95%CI: 1.03-1.43). As the birth weight gain, the slope of the dose-response curve of breast cancer increased continuously and the curve of testicular cancer was U-shaped. (Pnonlinearity<0.001) In the MR study, seven cancers were included. Only invasive mucinous ovarian cancer was found to have casual effect on birth weight (OR: 0.62; 95%CI: 0.39-0.97) while other cancers did not. Conclusions There is a nonlinear dose-response relationship between birth weight and breast cancer and testicular cancer. And birth weight has a casual effect on invasive mucinous ovarian cancer. Tweetable abstract Birth weight is associated with cancer risk but affects it indirectly.