Children with pervasive developmental disorders (PDD) are impaired by early onset difficulties in reciprocal social interaction. It has been proposed that this social impairment is the result of a lack of Theory of Mind (ToM): the ability to attribute mental states (e.g., beliefs, desires, intentions, and emotions) to others and to use these in predicting and explaining the behavior of others (Baron-Cohen, 1995; Serra et al., 1995). Research has shown that this ability is seriously impaired in autistic children (Baron-Cohen, Leslie, & Frith, 1985), and moderately impaired in children with Pervasive Developmental Disorders, Not Otherwise Specified (PDD-NOS) (Baron-Cohen & Swettenham, 1997; Happe, 1995; Perner & Wimmer, 1985). This impairment may limit PDD-NOS children in their understanding of humor and irony, and restrict their ability to take into account the interests and knowledge of a partner in social conversation (Serra, Loth, van Geert, Hurkens, & Minderaa, 2002). The hypothesis that an underdeveloped ToM underlies social impairment in PDD-children has led to the development of social skill interventions targeted at improving ToM-abilities. Ozonoff and Miller (1995) were the first to develop a group-based social-skills training, that has been evaluated in a controlled study with five adolescents (mean age: 13.8 [SD: 0.24]; Full Scale IQ scores above 70) participating in treatment and four adolescents in a no-treatment control group. Within two weeks, post intervention meaningful change was found in the treatment group’s performance on four false belief tasks (MM SecondOrder Belief Attribution Task, Baron-Cohen, 1989; Overcoat Story, Bowler, 1992; Prisoner Story, Happe, 1994), and no improvement in the control sample. No changes, however, were demonstrated on parent and teacher ratings of social competence for either group. Steerneman, Jackson, Pelzer, and Muris (1996) developed a program for children with social handicaps, including children with PDD (Steerneman et al., 1996). The program is developed for children with a range of social problems, but its effectiveness has only been examined in a study with eight children (age 6–8 years) with relatively mild social anxiety and social aggression (Steerneman et al., 1996). No children with PDD were involved in this study. Full scale IQ’s varied between 88 and 112. Before the training, directly after completion and four months after completion, emotion recognition (Test of perception of emotion from facial expression & Test of perception of emotion from posture cues, Spence, 1980) and ToM (Sally and Anne-test, Wimmer & Perner, 1983; pencil and Smarties box-task, Perner & Wimmer, 1985) were assessed in the children. At the same times social skills as reported by teachers and parents were evaluated. Compared to controls 1 Department of Child and Adolescent Psychiatry, Academic Medical Centre, Amsterdam, The Netherlands. 2 de Bascule, Academic Centre for Child and Adolescent Psychiatry, Amsterdam, The Netherlands. 3 Correspondence should be addressed to: Carolien Gevers, AMC – Dept Child Adol Psychiat./de Bascule, P.O. Box 12474, 1100 AL, Amsterdam, The Netherlands; Tel: +31-20-6501500; E-mail: c.gevers@debascule.com