A method, client, server and system for multi-person audio-video interaction are disclosed in the present invention, which belong to the field of internet terminal. The method includes: obtaining channel configuration information corresponding to current channel from a server; loading a scene module corresponding to the current channel according to the obtained channel configuration information and pre-stored configuration information; obtaining and displaying pictures or videos corresponding to the current channel from the server; displaying the text information by a simulation object corresponding to the client in a user interaction region of the scene module; and sending the text information to the server, when the text information that triggered by the user in the user interaction region of the scene module is received. The present invention achieves that every simulation object can be used to display the text information in different space position of the user interaction region, thus making an audience realize well communication and interaction effect in a scene environment simulated by a virtual scene.