Objective To examine the hypothesis that hyperinsulinaemia is associated with the development of borderline hypertension or hypertension. Design Blood pressure status in non-obese normotensives (< 140/90 mmHg, n = 135) people were re-examined after 11 years after the baseline examination. Participants were selected from a 1981 population-based health examination and had a high blood glucose level or more than a trace of glucose in their urine. Out of 319 people recruited for further examination of glucose tolerance status, 135 normotensive participants with body mass index < 26 kg/m2 and without diabetes according to World Health Organization criteria were re-examined at the follow-up survey. Results Sixty-two (46%) out of 135 normotensive participants were hypertensive (defined as blood pressure ≥ 140/90 mmHg) or receiving antihypertensive medication (n = 8) at the follow-up survey. Significant associations between the development of hypertension and baseline parameters were observed for systolic and diastolic blood pressure, serum triglycerides, high-density lipoprotein (HDL)-cholesterol, fasting and 60 min post-load insulin levels, and the sum of insulin concentrations from fasting to 180 min after glucose challenge after adjustments for age and sex. Odds ratios (95% confidence intervals) for the future development of hypertension between the highest and the lowest tertiles of insulin levels were 4.06 (1.40–11.76) for fasting insulin, 4.25 (1.45–12.45) for 60 min post-glucose load insulin, and 3.88 (1.34–11.20) for the sum of insulin concentrations, after adjustment for age, sex, systolic blood pressure, body mass index and alcohol consumption. Further adjustments for serum triglycerides and serum creatinine did not affect the insulin–hypertension relationship. Conclusion The present study suggests that hyperinsulinemia is significantly related to the development of hypertension in non-obese and non-diabetic Japanese people.