Objectives: To analyse how psychosocial determinants of lifestyle changes targeted in the Greater Green Triangle Diabetes Prevention Project conducted in Southeast Australia in 2004-2006 predict changes in dietary behaviour and clinical risk factors. Methods: A longitudinal pre-test and post-test study design was used. The group program was completed by 237 people at high risk of type 2 diabetes. Associations between changes in the variables were examined by structural equation modelling using a path model in which changes in psychological determinants for lifestyle predicted changes in dietary behaviours (fat and fibre intake), which subsequently predicted changes in waist circumference and other clinical outcomes. Standardised regression weights are presented, with β = ± 0.1 and β = ± 0.3 representing small and medium associations, respectively. Results: Improvements in coping self-efficacy and planning predicted improvements in fat (β = - 0.15, p<0.05 and β = - 0.32, p<0.001, respectively) and fibre intake (β = 0.15, p<0.05 and β = 0.23, p<0.001, respectively) which in turn predicted improvements in waist circumference (β = 0.18, p<0.01 and β = - 0.16, p<0.05, respectively). Improvements in waist circumference predicted improvements in diastolic blood pressure (β = 0.13, p<0.05), HDL (β = - 0.16, p<0.05), triglycerides (β = 0.17, p<0.01), and fasting glucose (β = 0.15, p<0.05). Conclusions: Psychological changes predicted behaviour changes, resulting in 12-month biophysical changes. The findings support the theoretical basis of the interventions.