This study evaluated the efficacy of approach-avoidance training as an additional treatment for children and adolescents with obesity seeking inpatient treatment. Two hundred thirty-two participants (8-16 years, 53.9% girls) were randomly assigned either to multisession approach-avoidance (IG) or to placebo training (CG). As outcomes, cognitive biases post intervention, body mass index, eating behaviour, food intake, self-regulation, and weight-related quality of life were assessed, also at 6- and 12-month follow-up. Modification of approach-avoidance bias was observed, but lacked in transfer over sessions and in generalization to attention and association bias. After 6 months, the IG reported less "problematic" food consumption, higher self-regulation, and higher quality of life; effects did not persist until the 12-month follow-up; no significant interaction effects were observed regarding weight course. Despite there was no direct effect on weight course, approach-avoidance training seems to be associated with promising effects on important pillars for weight loss. Further research concerning clinical effectiveness is warranted.