In a Voice over Internet Protocol (VoIP) video communication method and apparatus, first and second VoIP-based terminals store video event indexes and images mapped thereto with respect to state information of a user and terminal, and establish a session using session establishment information defining the video event. The first terminal then transmits a data packet including the video event index, and the second terminal receives the data packet and outputs an image mapped to the video event index of the images stored in advance. Thus, by transmitting and receiving a predefined index alone without transmitting user and terminal states as a real time data-type image, the VoIP video communication method and apparatus prevent unnecessary waste of bandwidth in video communication.