The long-term intelligent machine condition monitoring system is essential in improving maintenance costs and decision-making. Triboelectric nanogenerator (TENG) has a great advantage in developing self-powered machine condition monitoring. The main issues preventing TENG for such applications are poor integration to machine components, limited operational range, and weak durability. In this work, durable non-contact TENG energy harvester adaptive is designed to mechanical shafts for harvesting rotational energy. The harvester is designed to automatically switch to the non-contact mode by using centrifugal force to avoid manual switching while operating over a wide range of speeds of 0–2000 rpm. The designed TENG generates a high output of up to 25 mW with excellent stability for >20 days of continuous operation and exhibits a high-power density of 286 W m−3. Moreover, a self-powered long-term continuous condition monitoring system is developed from a TENG sensor, energy harvester, and wireless module. The developed system successfully sends possible machine fault frequency every 74 min to the cloud and accessible anywhere. This is the only TENG design reported in the literature that can fully power a Wi-Fi module to send data. Hence, the result promises the practical application of the system in developing internet of things (IoT) in the Industry.