MEBREK, Samir OussamaHAMLILI, Heyem2025-07-072025-07-072025http://dspace.univ-temouchent.edu.dz/handle/123456789/6431The loss of personal items such as keys, wallets, and important documents is a widespread issue, often resulting in stress, wasted time, and potential security risks. While existing tracking solutions offer post-loss recovery features, they tend to be expensive, depend on internet connectivity, and are frequently limited by proprietary ecosystems. All of these limitations ultimately make them less practical for global use, especially in countries like Algeria. This research presents ‘‘SafeFind’’, a preventative tracking system utilizing Bluetooth Low Energy (BLE) technology to effectively reduce the chances of losing belongings occurring before it happens. The SafeFind tag continuously checks the distance between the user’s phone and an item with the use of a machine learning component, a Random Forest classifier trained on real world data of BLE signal features like RSSI (Received Signal Strength Indicator). And detects patterns that signify a potential item lost situation. The system uses light sleep and deep sleep modes to maximize the battery lifetime of the tag. SafeFind additionally allows for the manual activation of an alarm, safe zone creation and the ability to support multiple devices through a cross-platform mobile app developed with Flutter.enBluetooth Low Energy (BLE), RSSI, Proximity Detection, Random Forest Classifier, Machine Learning, Preventive Tracking, IoT Security,Thesis