Objective: To investigate associations between general practitioner vocational training location and subsequent practice location, including the effect of rural origin. Design and participants: Annual panel survey of GPs (from the MABEL study) who completed their vocational training and transitioned to independent practice, 2008-2014. Main outcome measures: Rural practice location in the 5 years after vocational registration for participants in four primary cohorts: (1) rural origin/rural training; (2) metropolitan origin/rural training; (3) rural origin/metropolitan training; and (4) metropolitan origin/metropolitan training. Results: During the study period, 610 doctors completed GP vocational training and commenced independent practice. 74-91% of rural origin/rural training cohort GPs remained in rural areas during their first 5 years after completing training, with 61-70% remaining in the same community. Conversely, 87-95% of metropolitan origin/metropolitan training cohort GPs remained in metropolitan areas. GPs from the other two cohorts initially remained in their training location type, but gradually moved towards their origin type. Generalised estimating equation logit models identified a highly significant association between rural training pathways and subsequent rural practice that was sustained for 5 years after vocational registration; it was substantially strengthened when combined with rural origin (cohort 2 v cohort 4: odds ratio [OR], 24; 95% CI, 13-43; cohort 1 v cohort 4: OR, 52; 95% CI, 24-111). Conclusion: This study provides new quantitative evidence of strong associations between rural GP vocational training location and subsequent rural practice location, which is strengthened when combined with rural origin. This evidence supports current government policy supporting rural GP vocational training and quotas for medical student selection based on rural origin.