Fenced reserves are widely used to conserve and manage African savanna elephants (Loxodonta africana) in South Africa. However, the effects of fencing and fence removal on elephant wellbeing, movement and ecosystem dynamics remain limited in understanding, with few studies incorporating these elements. We hypothesised that fence removal would increase elephant movement flexibility and range, reduce behavioural disturbance and increase sociality, reduce stress-related biomarkers and redistribute browsing pressure. We evaluated how a resident elephant population in a private reserve in the Eastern Cape, South Africa, responded to the removal of two internal fences by integrating focal behavioural observations, faecal glucocorticoid metabolite (fGCM) analyses, NDVI-based vegetation assessments and GPS collar data collected before and after fence removal. Fence removal increased home-range extent, seasonal mobility and inter-herd spatial overlap, with elephants incorporating newly accessible habitat at variable rates. Initial movement was towards lower density environments during the exploratory phase, with greater distribution in space utilisation over time. Behavioural responses indicated short-term increases in aggression, vigilance and disturbance-related behaviours immediately after fence removal, followed by stabilisation, reduced aggression and increased affiliative behaviour over time. Activity budgets shifted towards increased locomotion and reduced foraging, consistent with exploratory behaviour. fGCM concentrations declined significantly following fence removal and remained significant after accounting for demographic and environmental variables, suggesting reduced physiological stress. Vegetation dynamics were spatially heterogeneous yet comparable between areas of elephant presence and absence. This suggests that expanded access redistributed browsing pressure without causing disproportionate vegetation loss with vegetation change primarily driven by environmental variability and not elephant access. These findings suggest that fence removal can promote more natural movement, behavioural responses and physiological regulation, while maintaining vegetation dynamics consistent with background climatic variability. Fence removal may therefore provide an alternative or complementary approach to population management by alleviating spatial constraints and redistributing ecological pressure across the landscape.