<p>Human–elephant conflict causes substantial damages to people and elephants worldwide through crop damage, property loss, injuries and fatalities. As interactions increase, identifying environmental drivers of reported elephant occurrence is essential for developing long-term mitigation strategies. We surveyed 420 community members living in conflict areas across Kerala and Karnataka, in the southern Western Ghats of south India using opportunistic and snowball sampling. We predicted that reported elephant occurrence would be positively associated with higher rainfall, lower human population density, higher years of residence, and suitable land cover, and negatively associated with distance to protected areas and water bodies. We used generalized additive models to estimate the probability of reported elephant occurrence, and a Markov Chain model to project future occurrence in 2030. Our results partially supported these predictions. Reported elephant occurrence increased with more rainfall and was higher when closer to protected areas and water bodies, however, human population density and years of residence showed significant non-linear effects, indicating their influence on reported elephant occurrence is more complex than anticipated. Across both landscapes, relationships between environmental variables and reported elephant occurrence shifted from near-linear in 2022 to non-linear patterns in 2030, indicating growing unpredictability in elephant movement under changing conditions. Thematic analysis of community responses revealed key themes of elephant habitat changes such as: perceived habitat changes and dense forest, increased human–elephant interactions, and mitigation measures. Findings highlight the importance of habitat restoration, maintaining functional ecological corridors, and integrating local knowledge with environmental modeling to better understand conflict and support coexistence.</p>