Objectives: To build on preliminary research evidence that suggests Leventhal’s Self Regulation Model (SRM) of illness behaviour can be used to understand how people make sense of depression. In particular the study aims were to explore whether the SRM can be used to understand the coping behaviour of people with depression and to assess whether beliefs about depression are related to self-reported adherence to anti-depressant medication and change in severity of depression following a group psychological therapy intervention. Design: The study utilised a cross-sectional survey design, with a longitudinal follow-up of a subset of the sample attending a group based therapy to explore whether beliefs about depression are related to change in severity of depression over the course of the intervention. Participants: Adults aged 18 to 65 years referred to two primary care psychology and counselling services for help with depression, including depression co-occurring with anxiety. Clients with depression co-occurring with more complex mental health problems such as substance misuse and personality disorder were excluded. The final sample consisted of 8 client’s attending individual therapy sessions and 60 clients attending a 7-week cognitive behavioural group therapy intervention. Outcome Measures: A Personal Models of Depression Questionnaire (PMDQ), based on a modified version of the Illness Perception Questionnaire-Revised, was developed to assess participant’s beliefs about depression. Outcome measures included the Short-Form COPE, the Medication Adherence Scale (MARS) and the Beck Depression Inventory-Version II (BDI-II). Results: The PMDQ showed good levels of internal consistency. Analysis of correlations between variables showed that, consistent with the theoretical tenets of the SRM, there were logical relationships and inter-relationships between the dimensions of the personal models o f depression, even when controlling for severity of depression. Beliefs were also associated with both adaptive and maladaptive coping behaviours and with self-reported adherence to antidepressants. However, personal models were not related to change in BDI-II scores following a group therapy intervention. Conclusions: The results of the study build on preliminary research evidence that suggests the SRM provides a useful theoretical model in which to understand how people make sense of and cope with depression. Future research should utilise longitudinal designs to test the predictive value o f the model in more detail. The clinical utility of assessing personal models of depression is discussed.