Clinical studies that compare lente insulin and neutral protamine Hagedorn (NPH) insulin in diabetic dogs are lacking. This is a prospective, randomised, controlled clinical study aimed to compare the efficacy and safety of lente insulin and NPH insulin in diabetic dogs. Thirty client-owned, newly diagnosed diabetic dogs were included. Animals were randomised into two groups and received lente insulin or NPH insulin administered every 12 hours. Follow-up re-evaluations were done at 1, 2, 4, 6, 8 and 12 weeks. At each re-evaluation, a physical exam, blood glucose curve, and serum fructosamine concentrations were performed. At the end of the study, the median insulin dose per injection was 0.61 U/kg (range, 0.34–0.92 U/kg) and 0.49 U/kg (range, 0.23–0.68 U/kg) in the lente and NPH groups, respectively. There was a significant improvement of polyuria and polydipsia and glucose concentrations in both groups. At the end of the study, the glycaemic control was considered good in 9/15 (60 per cent) and 11/15 (73 per cent) in the lente and NPH groups, respectively. These differences were not significant. Lente insulin and NPH insulin were similarly effective in the treatment of dogs with diabetes mellitus.