In modern society, workload makes more and more pressure to any office workers that obliges them sitting long time on chairs, leading to many health risks. This study proposes a solution to integrate the IoT (Internet of Things) technology into an office chair, so called IoT-chair. An IoT module measures the movements of workers and send data to a computer via WiFi connection. That data includes the weight of worker and the 3-axis acceleration provided by sensors. A feedforward neural network is trained to estimate the health status. When the worker does not move continuously for more than 3 minutes, the IoT-chair considers that is an abnormal situation and send a SOS message to an assistant. In addition, the IoT-chair can also make audible remind when the worker sitting longer than 45 minutes. Experimental results on some scenarios show that the ability to remind of long sitting conditions reaches 100% accuracy, and the ability to detect and warn abnormal health conditions reaches 82% accuracy. Experiments also show that it is possible to complete this product for a wide range of applications.