This secondary data analysis examined: (a) longitudinal trajectories of physical activity, sedentary behaviors, and obesity, (b) predictors of health behaviors (physical activity and sedentary behaviors) across age, (c) longitudinal relationships between obesity and health behaviors, and (d) interactions between physical activity and sedentary behaviors on obesity. The sample consisted of 3,805 subjects aged 8 to 19 years, who enrolled in the Cardiovascular Health in Children and Youth (CHIC) study from rural North Carolina. Each subject was observed a maximum 4 time points. SAS Proc Mixed (9.1.3) was used for analyses. Physical activity decreased and obesity increased as subjects grew older. Hours of TV viewing and video games also decreased with age. A faster decrease in physical activity with age was associated with early and mid-puberty, girls, black youth, children from the highest family income group and children with obese parents. Black youth watched TV and played video games more than their white counterparts. Boys spent more time in video games than girls. While vigorous activity was negatively related to obesity, the significance disappeared after adding parental characteristics in the models. Parental obesity was a stronger predictor for child obesity than child physical activity. Video games, not TV and computer use, were positively related to obesity longitudinally. Of particular importance sedentary behaviors were not inversely related to physical activity. Significant interaction between physical activity and sedentary behaviors were found; while more computer use strengthened the beneficial effects of physical activity on lowering obesity, subjects with more video game play did not have the beneficial effects of physical activity on lowering obesity. This study highlighted a critical period for intervening in physical activity and obesity for young adolescents. Because physical activity and sedentary behaviors are not in the same dimension, evaluation of both health behaviors at an individual level will be needed for the assessment of risk for obesity. In particular, even though subjects are highly active, the subjects may have higher risk for obesity if the subjects spend greater time in video games. Since parental obesity is a strong predictor for obesity, parents should be factored into obesity interventions for youth.