Human-wildlife conflict (HWC), particularly due to elephant foraging in agricultural fields and near human settlements, poses a serious challenge to both rural livelihoods and wildlife conservation in the Dooars region of North Bengal. This study investigates the application of Internet of Things (IoT)-based systems for proactive conflict mitigation. We propose a multi-layered IoT architecture integrating sensor networks—including motion detectors, infrared cameras, and acoustic sensors—for real-time detection and tracking of elephants. Additionally, spatio-temporal data on elephant movement and foraging patterns were analyzed using machine learning to identify high-risk zones and predict future incursions. This approach supports the strategic deployment of deterrents and better resource planning. This paper proposes a multi-layer IoT architecture (motion sensors, thermal/ infrared cameras, acoustic sensors) and alert system to detect and deter wild elephants entering farmland in North Bengal’s Dooars region. A pilot deployment (10 IoT nodes, LoRaWAN connectivity) was monitored for 3 months, yielding 67 elephant detections (61 true positives, 4 false negatives, 93.4% accuracy) and a marked reduction in crop damage incidents (from 12 to 3 per month) and HEC reports. It may be concluded that the IoT system significantly reduced foraging incidents and has strong potential for scaling. Ultimately, the research aims to validate a smart, data-driven solution for reducing HWC, promoting coexistence, and supporting long-term conservation of elephants in the ecologically sensitive Dooars landscape.