Objective. To evaluate the relative contributions of personal models and depression to the prediction of multiple outcomes in diabetes. Design and methods. Participants (N= 111) in a randomized trial of an intervention to increase dietary self‐management for diabetes patients were assessed for individual differences on three components of personal models (Seriousness, Treatment Effectiveness and Control)and depression on two occasions over a 3‐month period. The predictive utility of personal models versus depression was evaluated for dietary, physiologic and quality‐of‐life outcomes, controlling for baseline levels. Results. Personal models and depression were relatively stable over 3 months (r= .56‐.70). Personal models were predictive of outcomes in each category (eating patterns,p<.01; glycosylated haemoglobin,p<.01; physical functioning,p<.05; mental health,p<.05), whereas depression was only predictive of the mental health component of quality of life (p<.01). Conclusions. Personal models appeared to be stronger and more consistent predictors of outcome than depression. The implications of these findings were discussed in terms of the self‐regulatory model and interventions to improve diabetes self