Abstract: In a URL retrievl system and a URL retrieval method, a user is not required to assume a keyword for information to be accessed and in which even when many URL are obtained through a search, the user need not to select desired URL therefrom. A proxy search server creates a search expression using information from a user terminal. A search engine server extracts URL according to the search expression from the search server. If the number of the extracted URL exceeds a predetermined value, the search engine server sends information indicating for each URL a hierarchic level to which the URL belongs via the proxy engine server to the terminal for user selection. From the extracted URL, the engine server obtains URL belonging to a hierarchic level selected by the user. If the number of the URL does not exceed a predetermined value, the engine server sends the URL as a retrieval result via the proxy engine server to the user terminal for user selection.