BACKGROUND:  Availability of parks is associated with higher levels of physical activity among children and adolescents. Few studies examine actual park use and park-based physical activity in these populations. PURPOSE:  This study examined associations among individual, park, and neighborhood environmental characteristics and children's and adolescent's park-based physical activity. METHODS:  Data were collected in 2007 on 2712 children in 20 randomly selected parks in Durham NC. The System for Observing Play and Recreation in Communities (SOPARC) provided measures of physical activity. Hierarchic regression analysis assessed associations among individual, park, and neighborhood environmental characteristics and children's park-based physical activity. Data were analyzed in 2010. RESULTS:  Of the 2712 children observed, 34.2% and 13.2% were engaged in walking or vigorous physical activity. Environmental features of parks were associated with activity levels whereas neighborhood characteristics were not. Physical activity was negatively associated with gender (girls) (p=0.003); presence of a parent (p<0.0001); presence of nonparental adult (p=0.006); and an interaction involving the 0–5 years age group and style of play (p=0.017). Higher level of physical activity was associated with presence of other active children (p<0.0001); courts (e.g., basketball); and an interaction between number of recreation facilities and formal activities (p=0.004).  CONCLUSIONS:  These social factors and design features should be considered in order to stimulate higher levels of park-based physical activity among children and adolescents.