An image processing apparatus which is capable of improving user's convenience in searching for documents once found by a previous search, again. A CPU of the apparatus accepts the entry of a search keyword from the user, and obtains user information for identifying the user. The CPU adds the input search keyword to an index information item contained in each of document data items in association with the user information, and stores the index information item having the search keyword added thereto in an HDD, together with the document data item.