Understanding the factors that influence youth physical activity and overweight are important because promoting physical activity and preventing overweight are national health goals. The majority of research on factors that influence physical activity is cross-sectional and uses self-report measures of physical activity. To overcome these potential limitations in causality (cross-sectional designs) and validity (subjectively measured physical activity), the current study employed a longitudinal design and used an objective measure of physical activity. The primary purpose of this study was to determine if physical self-perceptions change over 27 months and if baseline physical self-perceptions predict follow-up physical activity. The secondary purpose was to determine if baseline physical self-perceptions predict follow-up body mass index (BMI). Participants were recruited from two elementary schools (grades 3-6) in a suburban school district in the southwestern United States. Complete data was collected on 229 participants at baseline and followed-up 27 months later. Yamax Digiwalker SW-200 (Tokyo, Japan) and Walk4Life My Life Stepper 2525 pedometers (Plainfield, Illinois, USA) were used to assess physical activity at baseline and follow-up respectively. Height and weight measures were collected at baseline and follow-up to determine BMI. The Children and Youth version of the Physical Self-Perception Profile (CY-PSPP) was used to assess physical self-perceptions both at baseline and follow-up. The CY-PSPP measures sport competence (sport), body attractiveness (body), physical condition (conditioning), physical strength (strength), physical self-worth, and general self-worth. A mixed design multivariate analysis of variance (MANOVA) procedure was conducted to determine if children's physical self-perceptions change over 27 months. The MANOVA revealed a significant time main effect (p < .01, partial η ² = .16). Post hoc analysis revealed that children's perceptions of body attractiveness (p < .004), physical strength (p < .001), and physical self-worth (p < .001) all significantly decreased over 27 months. Two separate hierarchical multiple regression analyses were conducted to predict follow-up physical activity and follow-up body mass index respectively. Both overall regression models were statistically significant, and explained 31% and 27% of the total variance for follow-up physical activity and follow-up body mass index respectively. Two subdomains of the CY-PSPP (physical condition & body attractiveness) were significant contributors for both follow-up physical activity and follow-up body mass index after controlling for demographics and baseline physical activity or BMI. The information derived on factors that influence both physical activity and BMI may aid interventionists and practitioners in designing effective programs to promote physical activity and prevent overweight in youth.