Hyperinsulinaemia and insulin resistance have been proposed to be associated with mortality risk, and diet can modulate insulin response. However, whether dietary patterns with high insulinaemic potential are associated with mortality remains unknown. We prospectively examined the associations between hyperinsulinaemic diets and the risk of total and cause-specific mortality in a large nationally representative population. Dietary factors were assessed by 24-h recalls. Two empirical dietary indices for hyperinsulinaemia (EDIH) and insulin resistance (EDIR) were developed to identify food groups most predictive of biomarkers for hyperinsulinaemia (C-peptide and insulin) and insulin resistance (homoeostatic model assessment for insulin resistance), respectively. Deaths from date of the first dietary interview until 31 December 2015 were identified by the National Death Index. Multivariable hazard ratios (HR) and 95 % CI were calculated using Cox regression models. During a median follow-up of 7·8 years, 4904 deaths were documented among 40 074 participants. For EDIH, the multivariable-adjusted HR (comparing extreme quintiles) were 1·20 (95 % CI 1·09, 1·32, P-trend<0·001) for overall mortality and 1·41 (95 % CI 1·15, 1·74, P-trend = 0·002) for CVD mortality. Similar associations were observed for EDIR with HR of 1·18 (95 % CI 1·07, 1·29, P-trend < 0·001) for total and 1·35 (95 % CI 1·09, 1·67, P-trend = 0·005) for CVD mortality. After further adjustments for BMI and diabetes, these positive associations were somewhat attenuated. Our findings suggested that diets with higher insulinaemic potential are associated with increased risk of overall and CVD-specific mortality.