An inverse association between coffee consumption and the risk of oral, pharyngeal, and esophageal cancers has been suggested in case-control studies, but few results from prospective studies are available. Data from the Miyagi Cohort Study in Japan were used to clarify the association between coffee consumption and the risk of these cancers. Information about coffee consumption was obtained from self-administered food frequency questionnaires in 1990. Among 38,679 subjects aged 40–64 years with no previous history of cancer, 157 cases of oral, pharyngeal, and esophageal cancers were identified during 13.6 years of follow-up. Hazard ratios were estimated by the Cox proportional hazards regression model. The risk of oral, pharyngeal, and esophageal cancers was inversely associated with coffee consumption. The multivariate-adjusted hazard ratio of these cancers for ≥1 cups of coffee per day compared with no consumption was 0.51 (95% confidence interval: 0.33, 0.77). This inverse association was consistent regardless of sex and cancer site and was observed both for subjects who did not drink or smoke and for those who currently drank or smoked at baseline. In conclusion, coffee consumption was associated with a lower risk of oral, pharyngeal, and esophageal cancers, even in the group at high risk of these cancers.