With the popularity of computer and network technologies, the number of web sites have rapidly increased, which has significantly stimulated the development of search engines. In recent years, a considerable number of search engines have been developed. The problem of how to find information on the Internet has been replaced with the problems of where to find the search engines, what they are designed to retrieve, and how to make the best use of them. Metasearch engine is a technique to cope with these problems by selecting and interfacing with multiple remote search engines at one time. One important issue on metasearch is how accurately it can select and rank feasible remote search engines for user queries. In this paper, a ratio-based metaindex approach is proposed for selecting remote search engines. Based on the novel approach, an experimental metasearch engine has been implemented. Some experimental results have shown that the accuracy and efficiency of the novel approach are more desirable than the famous metasearch engine SavvySearch.