Children with attention-deficit/hyperactivity disorder (ADHD) often encounter problems in social interactions with peers and are confronted with peer rejection and social isolation. The most common approach to social problems in children is social skills training. This intervention concept represents a variable mixture of cognitive-behavioral intervention elements. In this article the outcome of social skills training (SST) for children with ADHD is reviewed. Four experimental SSTs are detected and analyzed for potential mediators and moderators of treatment efficacy. Candidate mediators (social cognitive skills, parenting style and medication-induced reduction of key symptoms) are discussed within an empirical and theoretical context. Candidate moderators (subtype, comorbidity, gender and age) are evaluated for their empirical support. It is argued that, although fragmented, there is ample evidence and knowledge to adapt the SST-paradigm towards the specific needs of children suffering from ADHD and to guide future research towards more effective, "well established" interventions.