An RTP-based method of labial synchronization in H.323 video conference system is described. This scheme strictly obeys the RTP protocol, but brings the audio and video data into mutual connection for transmission via the same media channel, to achieve the purpose of labial synchronization. Experiments show that the scheme has not only little influence on the image quality, but also achieves a very good synchronous player of audio and video. This method is easy for operation. It has little burden on the system and can be used widely.