Objective: Investigate effects of persistent sleep disturbances during early childhood over ADHD during the adolescence, and the potential attention-related executive functions mediating this effect. Methods: We used data from the 2004 Pelotas Birth Cohort. Children's Sleep disturbances were reported by their mothers at 12, 24, and 48 months of age, whereas the Test-of-Everyday-Attention-for-Children (TEA-Ch) and the Development and Well Being Assessment (DAWBA) were applied at 11 years of age to evaluate attention-related executive functions and ADHD, respectively. Persistent sleep problems were defined as reporting have two or more points of difficulty to sleep, nightmares, restless sleeps, and/or <10h/24h sleep duration. Logistic regression and mediation models were used, adjusting for maternal and child sociodemographic, behavior and health related variables. Results: The highest prevalence of adolescent ADHD (15.4%) was on the group who reported having nightmares at 2,4 and 6 years. In adjusted models, we observed an odd of ADHD in the adolescence 2.26 higher in those who reported persistent nightmares (CI95% 1.33, 4.01) compared to those reported transitory or no nightmares. Persistent difficulty to sleep (OR=1.74 CI95% 1.13, 2.66) and restless sleep (OR=1.80, CI95% 1.23, 2.64) during childhood also increased ADHD odds at 11 years. No indirect effect through attention related executive functions was found using mediating models. Discussion: Persistent early sleep disturbances may increase odds of ADHD among adolescents and could be consider as early marker of such disorder, specially nightmares problems. These effects were not mediated by attention-related executive functions. Nevertheless, we had 75% of cohort inception response.