Abstract Background Controlling postprandial hyperinsulinaemia is important in insulin dysregulated (ID) horses to reduce the risk of laminitis. Objectives To evaluate postprandial insulin responses of ID versus non‐insulin dysregulated (NID) horses to feedstuffs varying in nonstructural carbohydrate (NSC) and crude protein (CP). Study design Randomised crossover. Methods Eighteen adult mixed‐breed horses (13.3 ± 2.2 years; 621 ± 78.8 kg) were individually fed [~1 g/kg body weight (BW)] specific feedstuffs within two crossover studies. Eight ID and eight NID were used in Study A, and 11 ID and 5 NID in Study B. In Study A, all horses were randomly fed once: cracked corn (CC: ~74% NSC & ~9% CP), ration balancer with low protein (RB‐LP: ~15% NSC & ~17% CP), ration balancer with high protein (RB‐HP: ~14% NSC and ~37% CP) and 50:50 mixture of RB‐LP:RB‐HP (MIX‐P). In Study B, horses were randomly fed once: CC, RB‐HP, steam‐flaked corn (SF: ~73% NSC & ~10% CP), oat groats (OG: ~64% NSC & ~14% CP) and a low NSC pellet (L‐NSC: ~6% NSC & ~12% CP). Blood was collected for insulin determination [radioimmunoassay (RIA)] before and 30, 60, 75, 90, 105, 120, 150, 180, 210 and 240‐minute post‐feeding in Study A and at 60‐minute in Study B. Data were analysed via analysis of variance (ANOVA) for repeated measures after any required transformations. Results ID horses had significantly greater insulin responses (AUCi) than NID for all diets in both studies ( P < .001; ID 22 362 ± 10 298 µIU/mL/min & NID 6145 ± 1922 µIU/mL/min). No effect of diet on AUCi for NID ( P = .2), but in ID, the CC (32 000 ± 13 960 µIU/mL/min) AUCi was higher than RB‐LP ( P = .01; 18 977 ± 6731 µIU/mL/min). ID insulin (T60) was lower for the L‐NSC (57.8 ± 18.5 µIU/mL) versus all other diets ( P < .02; 160.1 ± 91.5 µIU/mL). Main limitations Small numbers of horses; no ponies. Conclusions NSC appears to be the main driver of the postprandial insulin response. ID horses respond disproportionately to feeding even small amounts of low/moderate NSC feedstuffs. Data on possible dietary thresholds for postprandial insulin responses cannot be extrapolated from NID horses.