Few studies have examined antiretroviral therapy adherence in Latin American children. Standardized behavioral measures were applied to a large cohort of HIV-infected children in Brazil, Mexico, and Peru to assess adherence to prescribed antiretroviral therapy doses during the three days prior to study visits, assess timing of last missed dose, and evaluate the ability of the adherence measures to predict viral suppression. Time trends in adherence were modeled using a generalized estimating equations approach to account for possible correlations in outcomes measured repeatedly in the same participants. Associations of adherence with HIV viral load were examined using linear regression. Mean enrollment age of the 380 participants was 5 years; 57.6% had undetectable' viral load (<400 copies/mL). At enrollment, 90.8% of participants were perfectly (100%) adherent, compared to 87.6% at the 6-month and 92.0% at the 12-month visit; the proportion with perfect adherence did not differ over time (p=0.1). Perfect adherence was associated with a higher probability of undetectable viral load at the 12-month visit (odds ratio=4.1, 95% confidence interval: 1.8–9.1; p<0.001), but not at enrollment or the 6-month visit (p>0.3). Last time missed any antiretroviral therapy dose was reported as "never" for 52.0% at enrollment, increasing to 60.7% and 65.9% at the 6- and 12-month visits, respectively (p<0.001 for test of trend). The proportion with undetectable viral load was higher among those who never missed a dose at enrollment and the 12-month visit (p≤0.005), but not at the 6-month visit (p=0.2). While antiretroviral therapy adherence measures utilized in this study showed some association with viral load for these Latin American children, they may not be adequate for reliably identifying non-adherence and consequently children at risk for viral resistance. Other strategies are needed to improve the evaluation of adherence in this population.