Childhood obesity treatment programs only result in moderate outcomes in the short term and do not reduce risk for future weight gain. Therefore, in the current study, Approach Avoidance Training (AAT) with motivational game elements will be added to an inpatient childhood obesity program with the aim of improving outcomes. Forty-one children (10–15 years) in the final months of an inpatient treatment program were randomised to either the AAT plus care-as-usual condition group or to a care-as-usual-only control group. During the 10 sessions, the children were trained to approach healthy food stimuli and to avoid unhealthy food stimuli. Treatment outcomes were child performances on tasks of AAT, implicit attitudes and attentional bias, self-report ratings on craving symptoms, and weight loss maintenance after leaving the clinic (12-week follow-up). Changes over time were not significantly different between conditions for the measures of automatic processes, craving, and weight loss maintenance. Possible accounts for the null findings, including sample size, influence of game elements, point of time in therapeutic process, limitations of the setting, and the control group are discussed. More research is needed before firm conclusions can be drawn about the clinical usefulness of CBM for weight control in childhood obesity.