Current recommendations call for children to be physically active 60 minutes each day, with bouts of continual activity lasting at least 15 minutes. Although the importance of physical activity (PA) for children is recognized, there is little empirical evidence concerning how activity levels relate to children's health. Several measurement difficulties are inherent in the examination of this relationship. One is the valid and reliable assessment of PA, and another is the separation of performance-based fitness indicators from those that reflect optimal health. The purpose of this study was to investigate the complex relationship between children's engagement in PA and health-related fitness indicators by employing a design that encompassed multiple indicators of these constructs. It was hypothesized that children with higher levels of cardiovascular endurance and lower (healthier) levels of BMI and percent body fat would have higher levels of PA. Participants were 827 middle school physical education students. Health-related fitness measures were body mass index (BMI), percent body fat estimated from skinfolds, and the Hoosier endurance shuttle run (Safrit, 1995) as an indicant of cardiovascular fitness. Engagement in physical activity was assessed using three data sources: a self-report questionnaire (PAQ-C, Crocker et al., 1997), intention to be active (Standage et al., 2003) and pedometer step counts during physical education classes. Simple correlations revealed a negative relationship between cardiovascular fitness and percent body fat (r = -.62), as well as BMI (r = - 59). BMI and percent body fat had a strong positive relationship (r = .78). Self-reported PA and intention to be active were positively associated (r = .48), but both of these measures were unrelated to pedometer counts. A canonical correlational analysis was used to assess the multivariate relationship between engagement variables and health-related fitness indicators. The first function of the canonical correlation was significant [observed correlation = .30, Wilks's lambda = .89, F (9,1210) = 6.36, p < .001]. Average steps taken during physical education and intention to engage were significant contributors to the canonical variate for engagement. Cardiovascular fitness was the most powerful influence among the health indicators, with BMI emerging as a significant contributor. These findings provide insight concerning the affect of PA on children's health-related fitness. They support the assertion that active children are leaner and demonstrate higher levels of cardiovascular fitness, but the relationships among the variables also illustrate that accurate assessment of children's PA is a challenge.Keyword(s): exercise/fitness/physical activity, health promotion, physical education PK-12