Disclosed are a rich-media provision system and a control method thereof. The rich-media provision system includes: a media server for receiving predetermined element media constituting rich-media from an external server and storing the received element media; a media transmission server for providing the element media in the form of streaming and tracing a scene being currently reproduced by a terminal when predetermined capture input is detected; and a rich-media terminal for receiving and reproducing the traced scene. By the rich-media provision system, it is possible to previously capture and store a specific part of rich-media contents that a user desires, so that the user can use the captured part of the rich-media contents anywhere and anytime if necessary.