Background: It is expected that people with diabetes, throughout their lives, integrate and initiate a range of behavioral, therapeutic or preventive actions, suggesting the confirmed risk of occurring a globalized noncompliance, deteriorating their quality of life and an exponential economic impact. Objectives: to determine adherence to prescribed therapeutic regimen; and identify sociodemographic, clinical and psychosocial variables that influence adherence to the therapy. Methods: Was conducted a study quantitative, cross-sectional, non-experimental, descriptive, correlational study, with a sample of 102 people with diabetes type 2, aged 40 to 85 years, mostly male (51.96%). The evaluation protocol includes: social-demographic and clinical questionnaire, Diabetes Self-care Scale, Questionnaire about the knowledge of Diabetes, Depression, Anxiety and Stress Scale. Also resorted to HbA1c to directly assess adherence. Results: It appears that there is no statistically significant association between socio-demographic variables, sex and age and adherence. Single individuals, residents in urban areas, pensioners and those with the 3rd stage of schooling or more, adhere better to treatment. The variables such as, blood glucose monitoring, specific diet fulfillment and knowledge, reveal a statistically significant effect on adherence (p < 0.05). Anxious, depressed and stressed individuals adhere less. Conclusion: The evidences underlines the urgent need to recognize the importance of measuring patient adherence to a diabetes treatment plan for the maintenance of glycemic control. We suggest the reinforcement of educational programs in people with type 2 diabetes in order to enhance greater adherence to self-care.