Habitat fragmentation is a major biodiversity threat. Linear infrastructures (e.g. roads) hamper the movement of individuals and cause non-natural mortality. Roadkill hotspots have been used to define priority areas for road effect mitigation, but data availability and reliability is an issue, particularly on wide spatial scales. Additionally, mitigating the whole infrastructure network is unfeasible. Expedite methods are required to address such challenges. We present the gDefrag package, a graph-based approach that builds on habitat value and accessibility after simplifying the landscape as a graph. Its advantages include not requiring roadkill or movement data, and providing effective methods to deliver reliable information, allowing landscape managers to address landscape fragmentation overall. gDefrag prioritizes roads which should be targeted first to defragment the landscape. The software includes a user-friendly manual and currently implements four prioritization criteria: habitat quality, maximum number of inter-habitat paths, overall landscape connectivity, and simultaneously larger and higher-quality habitats.