Epidemiologic studies that examine whether isoflavone consumption protects against breast cancer have yielded inconsistent results. The controversy focuses on the effects of the menopausal status and exposure dose of isoflavone. We aim to conduct a meta-analysis on the association between isoflavone intake and breast cancer risk by comprehensively assessing isoflavone exposure in the targeted populations. We searched PUBMED and EMBASE databases for case-control and cohort studies that assess the association between isoflavone intake and breast cancer risk. We extracted relative risks (RR) and odds ratios (OR) of different reported categories of isoflavone intake from each study. Fixed- or random-effects models were used to summarize dose-response data. Twenty-two studies were selected for the meta-analysis. Overall, the results showed that isoflavone reduced the breast cancer risk (a combined RR/OR of 0.68, 95% CI: 0.52-0.89) in Asian populations rather than Western populations (a combined RR/OR of 0.98, 95% CI: 0.87, 1.11) for the high-dose category. Further analysis showed that the intake of isoflavone in postmenopausal Asian women 0.46 (95% CI: 0.28-0.78) was better than premenopausal 0.63 (95% CI: 0.50-0.80) but similar in postmenopausal Western women 1.00 (95% CI: 0.98-1.02) and premenopausal 0.99 (95% CI: 0.87-1.12). Exposure to high isoflavone may be associated with a reduced breast cancer risk in Asian populations, especially in postmenopausal women. However, no significant difference in the studies of Western populations may be due to the low intake of isoflavone levels.