BACKGROUND: Neighborhood perceived/built environment and physical activity (PA) associations have been examined for adolescents around homes, but not surrounding schools. PURPOSE: The purpose of this paper is to examine if positive perceptions/built environment in neighborhoods surrounding schools predict PA among low-income, urban adolescent girls. METHODS: Measures include: minutes in moderate-vigorous PA (MVPA, ankle accelerometry), perceptions of the school environment (questionnaire), built environment (neighborhood audit). Analyses include multi-level models. RESULTS: Two hundred twenty-four sixth and seventh grade girls [mean(sd) age = 12.1(0.7) years] from 12 schools serving low-income, primarily African American communities; mean MVPA 35.4 min (mean days assessed = 5.8). Girls in schools with more positive perceptions of the neighborhood environment surrounding the school were less active (β = 7.2, p = 0.043). Having “places to go within walking distance” (perceptions) and number of food stores near school (built environment) positively relate to MVPA (β = 5.5, p = 0.042 and β = 0.59, p = 0.047). CONCLUSIONS: Among neighborhoods surrounding urban schools, positive perceptions do not predict PA; accessibility, via both perceived and built environment, support PA.