OBJECTIVES: Diabetes mellitus (DM) is a chronic disease considered a worldwide public health problem. DM can result in harm to the patient in the aspects clinical, social, economic and quality of life, given its potential of morbidity and mortality DM (SBD, 2003). Expenditures for financing costly dispensing drugs have shown continuous growth trend in the Brazilian National Health Service (SUS). This creates the need for rationalization and optimization of growing financial contributions of resources for health. The purpose of this study was to evaluate the cost-effectiveness of Glargine - insulin analogue versus NPH - insulin in the treatment of diabetes mellitus type I and II. METHODS: We developed a decision model using Markov method to perform the cost-effectiveness of Glargine and NPH in the treatment of diabetes mellitus type I and II. Calculations were performed for a time horizon of 10 to 20 years in annual cycles. The perspective is that of the SUS. Additionally analysis Cost-utility and sensitivity were performed. RESULTS: In 20 years (with 5% discount), the effectiveness was equal (as in 10 years) between Glargine and NPH insulin presenting 7.62 Avg. The total cost to NPH insulin was U$9,114.95 and the total cost of insulin Glargine was U$21,128.49 In 20 years (with 5% discount), the utility was similar (as in 10 years) between Glargine and NPH insulin presenting 9.06 QALY. The reason of utility cost was U$1,006.03 and U$2,331.98 respectively. According to sensitivity analysis for drug prices (minimum and maximum), NPH was considered alternative more cost-effective in all scenarios. Exchange R$2, 67 = U$ 1.00 (Jan/15) CONCLUSIONS: Considering the similarity of the results obtained from the analysis of effectiveness and utility in comparing Glargine versus NPH at MD treatment, it is concluded that the difference remains only in the costs of treatment: Glargine costs are higher.