Abstract: To solve the problem that the current text-driven virtual effects cannot meet the practical application, this paper used the speech synthesis technology to complete the virtual vision from the waveform conversion to the human voice; used the face to synthesize the video effect of video synchronization; and used the thin-plate spline transformation method to drive virtual human images so as to synthesize synchronized virtual human images for audio and video. The experimental results realize the effect of virtual image, which can be used in the application problem of poor video effect and asynchronous text-driven effect, and has the same application scene.