Adolescent substance use is an increasing problem in the United States, and some researchers posit a bidirectional relation between adolescent substance use and the personality trait of impulsivity (e.g., Quinn, Stappenbeck, & Fromme, 2011). Friend substance use has been shown to be a powerful predictor of adolescent substance use, with prior research suggesting a bidirectional relation between adolescent substance use and friend substance use (e.g., Simons-Morton & Chen, 2006). Extant literature has not tested the bidirectional relation between adolescent substance use and impulsivity with longitudinal data nor has it examined this relation while considering the bidirectional relation with the social context factor of friend substance use. Using three waves of longitudinal data, we tested if there was a bidirectional relation between adolescent substance use and impulsivity while also examining the influences of friend substance use. Participants were 131 adolescents (male = 55%, mean age = 13 years at Wave 1). We tested nested models and examined whether adding equality constraints degraded the model fit using a Wald test. Results of structural equation modeling indicated that, after controlling for baseline levels of substance use, impulsivity predicted adolescent and friend substance use over time, whereas adolescent and friend substance use did not predict impulsivity. Adolescents with substance using friends were likely to increase their own substance use. The findings imply that aiming at both improving adolescents' ability to regulate impulsivity and deterring associations with friends who are using substances is essential for prevention and intervention efforts against substance use development in adolescents.