1. Fencing is one of the commonest methods for mitigating human-wildlife conflicts. At the same time, fencing is considered to be of one of the most pressing emerging threats to conservation globally. Although fences act as barriers and eventually can cause population isolation and fragmentation, it is challenging to quantitatively predict the possible consequences fences have for wildlife. 2. Here, we model how fencing designed to mitigate human-elephant conflict (HEC) on the Borderlands between Kenya and Tanzania will affect functional connectivity and movement corridors for African elephants. Specifically, we (1) model functional landscape connectivity integrating natural and anthropogenic factors; (2) predict seasonal movement corridors used by elephants in non-protected areas; and (3) evaluate whether fencing in one area can potentially intensify human-wildlife conflicts elsewhere. 3. We used GPS movement and remote sensing data to develop monthly step-selection functions to model functional connectivity. For future scenarios, we used a currently ongoing fencing project designed for human-elephant conflict mitigation within the study area. We modelled movement corridors using least-cost path and circuit theory methods, evaluated their predictive power and quantified connectivity changes resulting from the planned fencing. 4. Our results suggest that fencing will not cause landscape fragmentation and will not change functional landscape connectivity dramatically. However, fencing will lead to a loss of connectivity locally and will increase the potential for HEC in new areas. We estimated that wetlands important for movement corridors will be more intensively used by the elephants, which may also cause problems of overgrazing. Seasonal analysis highlighted an increasing usage of non-protected lands in the dry season and equal importance of the pinch point wetlands for preserving overall function connectivity. 5. Synthesis and applications. Fencing is a solution to small-scale HEC problems, but will not solve the issue at a broader scale. Moreover, our results highlight that it may intensify the conflicts and overuse of habitat patches in other areas, thereby negating any conservation benefits. If fencing is employed on a broader scale, then it is imperative that corridors are integrated within the protected area network to ensure local connectivity of affected species.