Background and objectiveYoung offenders show high levels of substance use. Treatment programs within detention settings are less effective. Cognitive bias modification (CBM) is a promising supplement to substance use treatment. This study tests the effectiveness of CBM in young offenders to reduce cannabis and alcohol use, and delinquent recidivism. MethodA randomized controlled trial added CBM to treatment as usual (TAU), among 181 youth in juvenile detention centers. In a factorial design, participants were randomly assigned to either active- or sham-training for two varieties of CBM, targeting attentional-bias (AtB) and approach-bias (ApB) for their most used substance. Substance use was measured with the Alcohol and Cannabis Use Disorder Identification Tests. Delinquent recidivism was measured with the International Self-Report Delinquency (ISRD) survey. ResultsAt pretest, participants showed AtB but no ApB for both substances. For alcohol, a decrease was found in AtB in the active-training group. For cannabis, a decrease was found in AtB for both active- and sham-training groups. Regardless of condition, no effects were found on substance use or ISRD scores at follow-up. LimitationsThe sample is judicial, not clinical, as is the setting. TAU and participant goals are not necessarily substance related. ConclusionsYoung offenders show a significant attentional-bias towards substance cues. CBM changed attentional-biases but not substance use. Combining CBM with a motivational intervention is advised. Follow-up research should better integrate CBM with running treatment programs. New developments regarding CBM task design could be used that link training better to treatment.