A method and system for handling multimedia calls is disclosed in which an IP multimedia terminal initiates a multimedia call request to a 3G terminal via a video inter-work device; the video inter-work device negotiates with 3G network and sets up a speech bearer between the video inter-work device and the 3G terminal when the 3G network is unable to support the multimedia call, then sets up a logical speech channel between the video inter-work device and the IP multimedia terminal, and makes the multimedia call fallback to a speech call. In accordance with the disclosed handling method and system provided, a multimedia call will fallback to a speech call when the 3G terminal does not support the multimedia call initiated by an IP multimedia terminal such that an H.324M video service is made more acceptable to users and complexity in using the service is reduced while improving user experience.