Objective Positional obstructive sleep apnea (POSA)—defined as obstructive sleep apnea twice as severe supine than nonsupine—may offer clues to the underlying pattern of upper airway collapse in children. We compared drug-induced sleep endoscopy (DISE) findings in children with and without POSA. We hypothesized that children with POSA would have significantly higher obstruction at the gravity-dependent palate and tongue base but not at the adenoid, lateral wall, or supraglottis. Study Design Retrospective case series. Setting Tertiary pediatric hospital. Subjects and Methods We included children aged 1 to 12 years with obstructive sleep apnea diagnosed by polysomnography who underwent DISE from July 2014 to February 2019. Scores were dichotomized as ≥50% obstruction (Chan-Parikh 2 or 3) vs <50% obstruction (Chan-Parikh 0 or 1). Results Of 99 children included, 32 (32%) had POSA and 67 (68%) did not. Children with POSA did not differ from children without POSA in age, overall apnea-hypopnea index, sex, race, syndromic diagnoses, obesity, or history of adenotonsillectomy. In logistic regression models, odds of ≥50% obstruction were significantly higher at the tongue base (odds ratio, 2.77; 95% CI, 1.04-7.39) after adjustment for age, sex, obesity, previous adenotonsillectomy, and syndrome. No difference was noted at the adenoid, velum, lateral wall, or supraglottis. Conclusion POSA was associated with higher odds of obstruction on DISE at the tongue base but not at other levels.