Electromyography (EMG) has been widely used for detecting a person’s hand poses and remote control applications. However, the traditional EMG-based control systems have limitations such as short controlling range, and supporting the limited number of devices. There is a need for a more advanced system that can deal with the limitations while maintaining a high quality of services such as high accuracy level and controlling complex devices. Hence, we present a real-time and remote control Internet-of-Things system using EMG signals together with motion-related signals such as acceleration and angular velocity. A user wearing the Myo-band at his/her arm can remotely control devices via 8 different hand gestures. The entire system was implemented and tested via two use cases of home assistant and robot arm control. The results show that the presented system could achieve a high level of accuracy e.g., 100% accuracy for simple control and 90% accuracy for complex cases. This system can be a potential approach for smart home controlling and assisting disabled people.