Aim This study aimed to assess whether a higher insulin response increased the long-term risk of mortality in a non-diabetic population. Methods A total of 446 people with normal glucose tolerance (NGT) or impaired glucose tolerance (IGT) who participated in the Da Qing Diabetes Study, were stratified into quartiles subgroups according to their baseline insulin area under the curve (AUC) during oral glucose tolerance test, defined as Q1, Q2, Q3 and Q4. The participants were followed from 1986 to 2016 to assess the risk of death in association with the magnitude of post-load insulin response. Results Over 30 years, the rates of all cause death were 9.94, 14.81, 15.02, and 17.58 per 1000 person-years across the four groups respectively. The rate for cardiovascular disease (CVD) death was 5.14, 6.50, 6.80 and 10.47 per 1000 person-years. Compared with Q1, the risk of all-cause death was significantly higher in participants in Q4 (HR = 2.14, 95% CI 1.34-3.42), Q3 (HR = 1.94, 95% CI 1.20-3.14), and Q2 group (HR = 1.70, 95% CI 1.06-2.74). In the Fine-Gray model with non-CVD death as competing risk, the increased insulin AUC were also significantly associated with the CVD death (Q4 vs Q1, HR = 2.04, 95% CI 1.10-3.79). In the fractional polynomial regression analysis, a nonlinear association between insulin AUC and all-cause and CVD death was demonstrated. In addition, insulin AUC was associated with a progressively higher risk of all-cause death and CVD death (fractional power 3, P < 0.001). Conclusion A higher post-load insulin response was significantly associated with a long-term increased risk of all-cause and CVD deaths in the Chinese non-diabetic population. It suggests that people featured by this phenotype is a potential important target for further intervention.