The current study examined multilevel risk factors and developmental assets on longitudinal trajectories of aggressive behavior in a diverse sample of rural adolescents. Using ecological and social capital theories, we explored the impact of positive and negative proximal processes, social capital, and contextual characteristics (i.e., school and neighborhood) on adolescent aggression. Data came from the Rural Adaptation Project, which is a 5-year longitudinal panel study of more than 4,000 middle and high school students from 40 public schools in two rural, low income counties in North Carolina. A three-level HLM model (N = 4,056 at Wave 1, 4,251 at Wave 2, and 4,256 at Wave 3) was estimated to predict factors affecting the change trajectories of aggression. Results indicated that negative proximal processes in the form of parent-adolescent conflict, friend rejection, peer pressure, delinquent friends, and school hassles were significant predictors of aggression. In addition, social capital in the form of ethnic identity, religious orientation, and school satisfaction served as buffers against aggression. Negative proximal processes were more salient predictors than positive proximal processes. School and neighborhood characteristics had a minimal impact on aggression. Overall, rates of aggression did not change significantly over the 3-year study window. Findings highlight the need to intervene in order to decrease negative interactions in the peer and parent domains.