The invention provides an embedded Internet kara OK entertainment device and a method for controlling sound and images thereof, which relate to a household kara OK system. The prior art has the defects of limited song source, user experience homogenization, general audio-visual effect and the like. The entertainment device comprises an embedded processor, an audio processing system, a video coderand a memory. A microphone is connected with the embedded processor through the audio processing system, and the embedded processor is also connected with an Internet controller. When the entertainment device is used, the embedded processor processes a media file in real time to form a dynamic background and scrolling captions and output video signals; and at the same time, the decoded music audios and the microphone reverb audios are mixed and amplified to output audio signals. Because the entertainment device is connected with a remote server through the Internet, when a song needed by a singer is not available in the local memory of the device, a media file of the corresponding song is downloaded from the remote server through the Internet. The entertainment device increases song sources, entertainment and interactivity of the kara OK system and improves user experience and audio-visual effect.