Background: Hyperlipidaemia is considered a cause of other diseases that are clinically important and potentially life threatening. Combination of pea and barley as exclusive starch sources is known to interfere with glycemic control in diabetic dogs, but their effect on lipid profile of hiperlipidaemic dogs is yet to be evaluated. Twelve adult diabetic dogs were fed three dry extruded diets with different starch sources and different fat levels: peas and barley (PB), maize (Mi), and peas, barley and rice (Ba) with 15.7, 15.6 and 9.0% of their dry matter as fat, respectively. Plasmatic cholesterol and triglycerides concentration curves over 10 h were obtained after 60 days on each diet and with the same NPH insulin dose. ANOVA test or Friedman test were used to compare the dietary effects on triglycerides and cholesterol variables among the diets. Results: Dogs presented lower mean (p = 0.05), fasting (p = 0.03), and time 8-h postprandial (p = 0.05) triglyceridemia after PB diet period than Ba diet period and time 4-h postprandial (p = 0.02) lower after PB than Mi diet. Cholesterolemia mean, minimum, maximum, area under the cholesterol curve and times points: 2, 4, 6, 8 and 10-h postprandial, had lower values after PB ingestion in comparison to Mi, without difference to Ba diet. Conclusion: Inclusion of pea and barley, as exclusive starch sources, in therapeutic diets for diabetic dogs can minimize plasmatic triglycerides and cholesterol concentration at fasting and at different postprandial time, compared to the maize diet or diet with lesser fat content.