Many young children appear to have skills sufficient to engage in basic elements of cognitive behaviour therapy (CBT). Previous research has, however, typically used children from non-clinical populations. It is important to assess children with mental health problems on cognitive skills relevant to CBT and to compare their performance to children who are not identified as having mental health difficulties. In this study 193 6 and 7 year old children were assessed using a thought-feeling-behaviour discrimination task [Quakley et al. Behav. Res. Therapy 42 (2004) 343] and a brief IQ test (the WASI). Children were assigned to groups (at risk, borderline, low risk) according to ratings of their mental health made by their teachers and parents on the Strengths and Difficulties Questionnaire [Goodman, J. Am. Acad. Child Adolescent Psych. 40 (2001) 1337]. After controlling for IQ, children ‘at risk’ of mental health problems performed significantly less well than children with a ‘low risk’ of mental health problems. Before receiving CBT, children’s meta-cognitive development should be assessed and additional help provided to those with meta-cognitive difficulties.