Although disadvantaged youth are more likely to be victimized at school, victimization only partly explains their decreased feelings of safety at school. We applied a socioecological approach to test the hypotheses that the experience of poverty is associated with decreased feelings of safety at school, and that residential neighborhood features partly mediate the relationship between poverty and feeling less safe at school. This study draws on the Québec Longitudinal Study of Child Development (QLSCD) which began in 1998 with a representative population-based cohort of 2,120 5-month old infants (49.1% female) and their primary caregiver. The study also includes measures of ego-centred residential neighborhood exposures (based on a 500 m circular buffer zone surrounding the family's residential postal code) derived from a spatial data infrastructure. We used latent growth modeling to estimate youth's family poverty trajectory from age 5 months to 13 years, and structural equation modeling to test our hypotheses. The results suggest that youth experiencing chronic and later-childhood poverty felt less safe at school in part because they lived in neighborhoods that their parents described as being disorderly (e.g., demarked by the presence of garbage, drug use and groups of trouble-makers). These neighborhoods also tended to have less greenery (e.g., trees, parks) and more lone-parent households. Neighborhood features did not help explain the relationship between early-childhood poverty and feeling less safe at school. The findings suggest that targeting residential neighborhood features such as greenery and disorder could improve youth's felt safety at school, particularly for those experiencing chronic and later-childhood poverty.