This study examined the association between magnesium intake and type II diabetes risk among Japanese adults. Participants were 25 872 men and 33 919 women aged 45-75 years who had no history of diabetes. Magnesium intake was ascertained using a 147-item food frequency questionnaire. Odds ratio of self-reported physician-diagnosed type II diabetes over 5 years was estimated using logistic regression analysis. A total of 1114 new cases of type II diabetes were self-reported. Magnesium intake was not significantly associated with type II diabetes in either men or women. The multivariable-adjusted odds ratios (95% confidence intervals) of type II diabetes for the highest versus lowest quintile of magnesium intake were 0.86 (0.63-1.16) and 0.92 (0.66-1.28) for men and women, respectively. Although a small effect cannot be excluded in men, magnesium intake may not be appreciably associated with risk of type II diabetes for Japanese adults.