A method for exchanging media description information between user agents using a session initiation protocol (SIP) comprises the steps of: providing information on an OS and sound module of a first user agent so that the first user agent generates, in a media description field of a session description protocol (SDP), an SIP message including the information on the OS and sound module of the first user agent, and sends it to a second user agent; and setting a processing format for voice data of the second user agent so that the second user agent determines the voice data format supportable by the first user agent from the information on the OS and sound module of the first user agent, compares it with the voice data format supportable by the second user agent, and sets a format commonly supportable by the first and second user agents as the voice data format of the second user agent.