Previous studies have shown that the incidence of diabetes mellitus (DM) has increased in human immunodeficiency virus (HIV)-infected patients with long-term exposure to highly active antiretroviral therapy (HAART). However, the factors associated with DM among HIV-infected patients in Asia remain unclear in the HAART era.A nationwide cohort studyData from Taiwan's National Health Insurance Research Database (NHIRD) between 2000 and 2010 were used to investigate the incidence of and factors associated with DM among HIV-infected patients. Propensity score matching was conducted to match 4797 patients receiving HAART (HAART cohort) with 4797 patients not receiving HAART (non-HAART cohort). HAART use was treated as a time-dependent variable in a Cox regression model.HAART cohort had a significantly higher 10-year incidence of DM (7.16%; 95% confidence interval [CI], 4.30%-10.03%) than non-HAART cohort (2.24%; 95% CI, 1.28%-3.20%) (P < .001). After adjusting for age, gender, and comorbidities, receiving HAART was associated with an increased incidence of DM, with a subdistribution hazard ratio (sHR) of 2.39 (95% CI, 1.65-3.45). Hypertension (sHR = 5.27; 95% CI, 3.21-8.65), gout (sHR = 2.39; 95% CI, 1.38-4.16), and hepatitis C virus (HCV) infection (sHR = 2.43; 95% CI, 1.28-4.61) were significantly associated with a higher risk of DM. Sensitivity analyses showed exposure to HAART remained significantly associated with an increased risk of DM, particularly in those without pre-existing hypertension, gout, or HCV infection.Exposure to HAART increased the risk of DM in HIV-infected Taiwanese patients, particularly in those without pre-existing hypertension, gout, or HCV infection.