In IoT network, a large amount of data is generated and transferred from the sources to destinations every day. The increasing data volume and continuous data transmission bring challenges to existing IoT system, such as the increased energy consumption and bandwidth usage, especially for battery-operated or cellular-connected IoT devices. Despite many researchers are working on these challenges, most of them are focused on the backend of IoT system, little attention has been given to the front-end data in the data reduction of IoT system. Therefore, we propose a users' behavior-aware data reduction framework for IoT, with the aim to provide a novel approach to data reduction in IoT. Consequently, our proposed methods can achieve a reduction of the amount of accumulated data transferred up to 65. © 2024 IEEE.