- Conventional video games are controlled by players via physical game controllers. A new emerging trend for game control in recent years is to use intelligent sensor based interface technology to control games directly by players' natural gestures of hand, arm and body movements. This paper presents a framework of sensor based interface for video games. In this system, a single TOF 3D camera is used as the sensor, which provides simultaneously range image and grayscale image. The two types of data are processed in parallel and then integrated to translate players' target gestures into game control parameters. The framework includes the following functionalities: 1) motion feature extraction and object segmentation; 2) gesture representation and modeling; 3) target gesture recognition, and 4) game parameter generation. A video Dart game is used as a test bed sample.