The relationship between sleep duration and obesity in adolescents is inconclusive. This may stem from a more complex relationship between sleep and obesity than previously considered. Shifts toward evening preferences, later sleep-wake times and irregular sleep-wake patterns are typical during adolescence but their relationship to body mass index (BMI) has been relatively unexplored. This cross-sectional study examined associations between sleep duration, midpoint of sleep and social jet lag (estimated from 7 days of continuous actigraphy monitoring), and morningness/eveningness with BMIs (BMI z-scores) and waist-to-height ratios in 14-17-year-old adolescents. Seventy participants were recruited from ninth and tenth grades at a public high school. Participants' characteristics were as follows: 74% female, 75% post-pubertal, 36% Hispanic, 38% White, 22% Black, 4% Asian and 64% free/reduced lunch participants with a mean age of 15.5 (SD, 0.7). Forty-one percent of the participants were obese (BMI ≥ 95th percentile); 54% were abdominally obese (waist-to-height ratio ≥ 0.5). Multivariable general linear models were used to estimate the association between the independent variables (school night sleep duration, free night sleep duration, midpoint of sleep (corrected), social jet lag and morningness/eveningness) and the dependent variables (BMI z-scores and waist-to-height ratios). Social jet lag is positively associated with BMI z-scores (p < 0.01) and waist-to-height ratios (p = 0.01). Midpoint of sleep (corrected) is positively associated with waist-to-height ratios (p = 0.01). After adjusting for social jet lag, school night sleep duration was not associated with waist-to-height ratios or BMI z-scores. Morningness/eveningness did not moderate the association between sleep duration and BMI z-scores. Findings from this study suggest that chronobiological approaches to preventing and treating obesity may be important for accelerating progress in reducing obesity rates in adolescents.