The source of starch may interfere with glycaemic control in dogs, but few studies have evaluated these aspects in diabetic dogs. This study compared the effects of two isonutrient diets with different starch sources, peas and barley (PB) v. maize (Mi), on diabetic dogs. The Mi diet was processed in order to generate a lower starch gelatinisation index. In all, fifteen adult diabetic dogs without other conditions were included. The animals were fed two dry extruded rations with moderate levels of fat and starch and high levels of protein and fibre using a random, double-blind cross-over design. Glycaemic curves over 48 h were developed via continuous glucose monitoring after 60 d on each diet and with the same neutral protamine Hagedorn (NPH) insulin dosage. The following were compared: fasting, mean, maximum and minimum blood glucose, maximum and minimum glycaemia difference, glycaemic increment, area under the glycaemic curve, area under the glycaemic increment curve and serum fructosamine concentration. Paired t tests or Wilcoxon signed-rank tests were used to compare the amount of food and nutrients ingested and the dietary effects on glycaemic variables between the diets. Dogs fed the PB diet presented a lower average mean interstitial glucose (P=0·01), longer mean hypoglycaemic time (P<0·01), shorter mean hyperglycaemic time (P<0·01) and smaller difference between maximum and minimum blood glucose levels (P=0·03). Thus, the processing applied to the Mi diet was not sufficient to achieve the same effects of PB on glycaemic control in diabetic dogs.