OBJECTIVE:Adolescent substance use has been associated with poorer neuropsychological functioning, but it is unclear if deficits predate or follow the onset of use. The goal of this prospective study was to understand how neuropsychological functioning during early adolescence could predict substance use by late adolescence. METHOD:At baseline, participants were 175 substance-use-naïve healthy 12- to 14-year-olds (41% female) recruited from local schools. Participants completed extensive interviews and neuropsychological tests. Each year, participants' substance use was assessed. By late adolescence (ages 17 to 18), 105 participants transitioned into substance use and 75 remained substance-naïve. Hierarchical linear regressions examined how baseline cognitive performance predicted subsequent substance use, controlling for common substance use risk factors (i.e., family history, externalizing behaviors, gender, pubertal development, and age). RESULTS:Poorer baseline performance on tests of cognitive inhibition-interference predicted higher follow-up peak drinks on an occasion (β = -.15; p < .001), more days of drinking (β = -.15; p < .001), and more marijuana use days (β = -.17; p < .001) by ages 17 to 18, above and beyond covariates. Performances on short-term memory, sustained attention, verbal learning and memory, visuospatial functioning, and spatial planning did not predict subsequent substance involvement (ps > .05). CONCLUSIONS:Compromised inhibitory functioning during early adolescence prior to the onset of substance use was related to more frequent and intense alcohol and marijuana use by late adolescence. Inhibition performance could help identify teens at risk for initiating heavy substance use during adolescence, and potentially could be modified to improve outcome.