As the rapid development of World Wide Web, there is t tremendous information "hiddened" in Hidden Web , and it s capacity is increasing rapidly. The information can only be accessed by t he query interfaces provided by Web database. The data in Hidden Web are obtained in the form of dynamic Web pages when users send a query. Due to the poor structure of web pages and the instability and large scale of Hidden Web, it is a very challenging task to integrate the abundant information automatically and use it effectively.Because of this hidden feature, hidden web is hard to crawl. It becomes a new direction in the field of information retrieval. In this paper the cause of formation and the feature of Hidden Web have been introduced. It also showed the similarities and differences of Hidden web crawler and traditional crawler by contrast. Through that, it analyzed the key technique of designing a Hidden Web crawler, and the step when a crawler crawling the web page.In this paper a new method of focused hidden web information retrieval is proposed. It presents a generic operational model of the hidden web information retrieval and describes the key techniques. It also introduces a focused crawling technique and a new heuristic query selection algorithm which designed by this paper. Based on those techniques, the crawling is more efficient and more accurate. The result of experiment indicates that the new solution is much better than the old one.