Non-organic recurrent abdominal pain (RAP) is a common experience among children, but is rarely attributable to a physical cause. Diagnostic processes associated with non-organic RAP are often expensive and characterised by inconclusive findings. The long-term prognosis, for untreated cases, is marked by pervasive symptomatology. It has been proposed that the presentation of RAP can be differentiated from organic abdominal pain by the presence of psychological disturbance in both the child and its family. However, the evidence for a direct causal link with these factors is equivocal. This review examines this evidence and suggests a psychological model-building approach to understanding the specific clinical features of RAP. The general principles of psychological model building are outlined and the literature considered in terms of antecedent, mediator and moderator variables. A psychological model is proposed based on this analysis and its clinical and research implications are discussed.