PURPOSE: The aim of this study was to compare the impact of fermented and unfermented soy intake, based on the following soy-derived products: tofu, soymilk, natto, and miso, on the risk of liver cancer among Japanese adults. METHODS: 75,089 Participants of the Japan Public Health Center-based Prospective Study (JPHC Study) were followed from the time of the 5-year follow-up questionnaire until the end of 2012-2013. Subjects with available data on hepatitis B virus (HBV) and hepatitis C virus (HCV) infection status from blood samples (n = 14,016) and those who were anti-HCV antibody (anti-HCV) or hepatitis B virus antigen (HBsAg) positive (n = 1033) were also analyzed separately. Cox proportional hazard models were employed to calculate hazard ratios (HRs) and 95% confidence intervals (95% CIs). RESULTS: During 1,145,453 person-years, 534 newly diagnosed cases of liver cancer were identified in the JPHC Study. For miso intake among men, the multivariate-adjusted highest versus lowest quartile HR was 0.65 (95% CI, 0.48-0.89); p for trend = 0.006. Results were similar in those who were anti-HCV or HBsAg positive, 0.24 (0.08-0.70); p for trend = 0.004 highest versus lowest tertile. For the sub-analysis among only participants with known hepatitis infection status and HCV and HBsAg adjustment, a similar association was observed. In the multivariate complete cohort analysis, among women, the highest intake of fried tofu was inversely associated with the risk of liver cancer, HR = 0.45 (0.26-0.80); p for trend = 0.014. CONCLUSIONS: We observed no association between total soy intake, fermented and unfermented, and risk of liver cancer, and only an inverse association between miso intake and liver cancer among men.