Attention bias for pain-related information is theorised to maintain chronic pain, indicating that changing this bias could improve pain-related outcomes. Modifying attention biases in adolescents, when chronic pain often first emerges, may be particularly beneficial. We report here a randomized, placebo-controlled, parallel-group trial of attention bias modification (ABM) training in adolescents with chronic noncancer pain. Adolescent patients (N = 66) were randomly assigned to complete multiple sessions of dot-probe ABM training (N = 23), placebo training (N = 22), or no training (waitlist; N = 21) across a period of 4 weeks. Patients completed all assessments at a hospital-based pediatric pain clinic and completed all training at home. We examined the relative effects of ABM on attention bias and attention control, as well as pain symptomatology (primary outcome), pain catastrophizing, anxiety and depression symptoms, and functional disability (secondary outcomes) immediately after training and 3 months later. We found no evidence that ABM changed attention bias or attention control in comparison with placebo training or no training. We also found that pain and pain-related outcomes were no different for those undergoing ABM compared with placebo training or no training when tested immediately after training or 3 months later. Overall, we found no evidence to support the efficacy of dot-probe ABM for improving pain-related outcomes in adolescents with chronic pain. This study was registered on the NIHR Clinical Research Network Portfolio in August 2014 (UK Clinical Trials Gateway: CPMS 17251) and funded by a Research Training Fellowship awarded to Lauren Heathcote by Action Medical Research for Children.