We searched four databases (Cochrane Library, PubMed, ScienceDirect, and Web of Science) for studies published before August 2023, screened eligible studies (inclusion and exclusion criteria based on population, exposure, comparator, and outcome framework), and collected data on exposure indicators and disease risks. We applied random-effects models of meta-analysis to estimate the pooled effects and meta-regression to assess the dose-response relationship (radiation dose rate for workers and distance for residents). We identified 47 studies, 13 with worker and 34 with resident samples, covering 175 nuclear power plants from 17 countries, encompassing samples of 480,623 workers and 7,530,886 residents. Workers had a significantly lower risk for all-cancer and a significantly higher risk for mesothelioma. Residents had significantly higher risks for all-cancer, thyroid cancer, and leukemia. Notably, children under 5 years old showed the highest risk for all-cancer. Our meta-regression showed a significantly positive dose-response relationship between cumulative dose of radiation exposure and risk for circulatory disease among workers. Our findings demonstrated higher risks for mesothelioma for workers and all-cancer, thyroid cancer, and leukemia for residents exposed to low-dose radiation from nuclear power plants. Some included studies did not adjust for cancer risk confounders, which could overestimate the association between radiation exposure and cancer risk and increase the risk of bias.