Cross-sectional studies have demonstrated that the brain-derived neurotrophic factor (BDNF) Val66Met single nucleotide polymorphism moderates the association between exposure to negative life events and depression outcomes. Yet, it is currently unclear whether this moderating effect is applicable to positive life events and if the moderating effect is stable over time. To address these gaps in the literature we examined clinical and BDNF genotypic data from a five-year prospective cohort of 310 primary care attendees. Primary care attendees were selected based on existence of depressive symptoms at screening. Depressive symptoms were assessed at baseline and annually for five years post-baseline using the Primary Care Evaluation of Mental Disorders Patient Health Questionnaire-9 (PHQ-9). Linear mixed models assessed differences in depressive symptom severity over the five-year follow-up period by BDNF Val66Met and history of life events, both negative and positive. Analysis identified a novel three-way interaction between the BDNF Val66Met polymorphism, history of severe childhood abuse and time. Post hoc analysis stratified by time, showed a two-way interaction between Val66Met and severe childhood abuse at baseline that was not detectable at any other time point. An interaction between Val66Met and positive life events was not detected. Our longitudinal results suggest the BDNF Val66Met polymorphism moderates the depressive symptom severity experienced by those with a history of severe childhood abuse but does so in a time dependent manner. Our results further support the notion that gene-environment-depression interactions are dynamic and highlight the importance of longitudinal assessment of these interactions. Given these novel longitudinal findings; replication is required.