Literature reviewing revealed inconsistent results regarding associations of Body Mass Index (BMI) with physical activity (PA), TV watching, and sports skill among teenagers. Further, although FITNESSGRAM and The President's Challenge provide standards for BMI for girls and boys respectively, some studies did not separate girls and boys when BMI was addressed. The study was intended to examine associations of BMI with PA, TV watching, and sports skill for girls and boys respectively. One hundred ninety nine middle school students (108 girls and 91 boys) in a suburb area in the Northeast region completed a 24-event, previous day PA recall questionnaire (Sallis et al., 1996), which was reduced to six PA variables: (a) minute of PA, (b) minute of moderate to vigorous PA (MVPA), (c) MET score, (d) MVPA MET score, (e) weighted MET score, and (f) weighted MVPA MET score. Compendium of physical activities (Ainsworth et al., 2000) was used for the calculation of energy expenditure, and FITNESSGRAM BMI standards were employed for classifying girls and boys in terms of BMI. While girls and boys had a quite similar distribution in normal weight group (75% girls, n = 81; 63.74% boys, n = 58), Chi Square tests identified that boys had a significant larger portion than girls (30.77% boys, n = 28; 5.6% girls, n = 6) in overweight groups, with χ2 = 26.63, p < .001, and Cramér's V = .366. Pearson correlation indicated significant negative correlations (r = -.311 or stronger, p < .001) between girls' BMI and each of six PA variables (larger BMI goes with lower PA amount), and significant positive correlation (r = .342, p < .001) between girls' MBI and the time in TV watching, but failed to reveal any significant association for the same variables for boys with family Type I error controlled. As for the correlation between BMI and sports ability measured by a researcher-generated rating sheet, Spearman correlation detected significant negative relationship (larger BMI goes with lower sports ability) for both girls (r = -.291, p = .002) and boys (r = -.308, p = .003). The results indicate that, although overweight girls and boys tend to have lower sports ability and boys have a larger overweight portion than girls, overweight middle school girls are more likely to spend less time in PA and more time in TV watching. It is overweight girls who need more PA intervention.