Family relationships habitually shape the way traumatic events affect children's mental health in a context of war and violence, but research is scarce on the role these relationships play in the success of psychosocial interventions. This study is a secondary analysis of previously identified family system types that are based on attachment, parenting, and siblingship, and of the influence they have on the effectiveness of teaching recovery techniques (TRT). TRT is a psychosocial group intervention aimed at reducing children's mental health problems and increasing their psychosocial resources. We tested three hypotheses. First was the compensation hypothesis, which holds that children from families with negative relationships benefit a great deal from the TRT intervention. The second was the accumulative hypothesis, which maintains that children from families with negative relationships do not benefit from the intervention. The third, the buffering hypothesis, states that children from families with positive relationships benefit a great deal from the intervention. The family sample consists of 325 Palestinian mothers and fathers and one of their children (age 10--13). Children participated either in the TRT intervention or waiting-list control groups. Their self-reported post-traumatic stress symptoms, emotional and conduct problems, positive resources, and prosocial behavior were assessed at baseline, three months post-intervention, and at a six-month follow-up. We found that family type was significantly associated with TRT effectiveness, which supports the compensation and buffering hypotheses. Children with insecure and negative family relationships and those from families with discrepant perceptions of relationship quality showed a decline in emotional problems across the three assessments, and an increase in positive resources. Children from families with highly secure, positive relationships and those with moderately secure, neutral relationships showed increased positive resources and prosocial behavior in the control group as well. We argue that a family system approach can deepen understanding of the mechanisms of successful psychosocial interventions and, therefore, that family relations should be taken into account when tailoring such interventions for traumatized children.