Children in foster care are at high risk of experiencing mental health problems and tackling this issue is a key priority. Previous research suggests that the transition from primary to secondary school can be particularly challenging, as well-being declines and mental health problems increase in early adolescence. However, there is insufficient understanding of variations in the well-being and mental health of this group of children, and particularly the role played by their social interactions, relationships, and psychological attributes. This thesis includes three papers reporting on a programme of empirical research conducted to address this gap in knowledge and better understand the risk and protective factors, particularly in the peer context, for changes in mental health and well-being. The first paper focuses on current provision and reports the findings from a national survey of Virtual Schools that support the education of children in care. The second paper presents the findings of a longitudinal study with children not in care (aged 10-13 years), to test our conceptual model in the general population. This demonstrated that peer factors predict changes in mental health problems and well-being over and above parental and other adult support. The third paper presents findings from a longitudinal study of children in foster care (aged 10-14 years), to test these key pathways in our focus population. This revealed a pattern of differentiated links from peer and adult support to mental health and well-being, and identified self-efficacy as a key longitudinal predictor of change, especially when moderated by peer relationship quality. The thesis demonstrates the importance of supportive relationships with both adults and peers for the mental health and well-being of children in care. This has important implications for future work where social activities and relationship quality with peers should be considered as potential protective factors, especially in school settings.