The purpose of the current study was to identify predictors of delay discounting among adolescents receiving treatment for marijuana abuse or dependence, and to test delay discounting as a predictor of treatment outcome. Participants for this study were 165 adolescents (88% male) between the ages of 12 and 18 (mean age = 15.8 years; standard deviation = 1.3 years) who enrolled in a clinical trial comparing three behavioral treatments for adolescent marijuana abuse or dependence. Participants completed a delay discounting task at treatment onset for $100 and $1,000 of hypothetical money and marijuana. Overall, smaller magnitude rewards were discounted more than larger magnitude rewards. Delay discounting rates were concurrently related to demographic variables (socioeconomic status, race). Delay discounting of $1,000 of money predicted during treatment abstinence outcomes among adolescent marijuana abusers, over and above the effects of type of treatment received. Teens who show higher levels of discounting of the future may be an important subgroup to identify at treatment onset. Youth with a greater tendency to discount the future may require different intervention strategies that address their impulsivity (e.g., targeting executive function or inhibitory control) and/or different schedules of reinforcement to address their degree of preference for immediate rewards.