Objective: We investigated the relation between fasting insulin (FI) and risk of cerebral infarction in a Japanese general population. Methods: The subjects were 2,610 men and women without past history of stroke or myocardial infarction and under treatment for diabetes, examined between 1992 and 1995 as part of the Jichi Medical School Cohort Study. The FI level was measured once at the baseline. Subjects were divided into quintiles by FI levels, and Cox’s proportional hazard model was used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for cerebral infarction. Results: During an average of 11.1 years of follow-up, 87 participants developed cerebral infarction. Crude incidence rates of FI quintiles 1-5 were 4.69, 2.35, 1.85, 2.77 and 3.30 per 1,000 person-years, respectively. The multivariate-adjusted HRs for cerebral infarction were 2.33 (95% CI, 1.10 – 4.96) in quintile 1 (Q1), 1.25 (95% CI, 0.55 – 2.84) in Q2, 1.68 (95% CI, 0.76 – 3.70) in Q4 and 2.06 (95% CI, 0.94 – 4.47) in Q5, using Q3 as the reference. Conclusions: The lowest FI level was associated with increased risk of cerebral infarction and the association between FI and risk of cerebral infarction appeared to be a U-shaped relationship.