Background: Many epidemiologic studies have explored the association between age at last birth (ALB) and the risk of epithelial ovarian cancer, but the results remain controversial. Methods: A literature search was performed in PubMed, Web of Science, China National Knowledge Infrastructure (CNKI) and WanFang Med Online for relevant articles published up to April 2019. Pooled relative risks (RRs) with 95% confidence intervals (CIs) were calculated using a random-effect model. Dose–response relationship was assessed by restricted cubic spline model. Results: Thirteen articles with 19,959 cases and 2,451,071 participants were included in our meta-analysis, and we found that ALB was negatively associated with epithelial ovarian cancer. The pooled RR (95% CI) of epithelial ovarian cancer for the highest versus the lowest stratification of ALB was 0.77 (0.65–0.91). Furthermore, significantly negative associations were shown in case–control studies (RR: 0.73; 95% CI: 0.60–0.88), studies conducted in North America (RR: 0.71; 95% CI: 0.60–0.84), studies with adjustment for parity (RR: 0.76; 95%CI: 0.63–0.93), studies with adjustment for tubal ligation (RR: 0.74; 95% CI: 0.58–0.94), in the subgroup analysis. In dose–response analysis, the risk of epithelial ovarian cancer decreased nonlinearly with the increase of ALB, and the negative results become significant when ALB was 22.5 years old. Conclusion: This meta-analysis suggested that ALB was negatively associated with the risk of epithelial ovarian cancer. The risk of epithelial ovarian cancer decreased gradually with the ALB for women.