Commercial serach engines, especially meta- search engines, were designed to locate the information by querying multiple conventional search engines and integrating the partial results generated by each search engine. A few meta- search engines that perform best for a particular query. The problem of finding the search engine that performs best for a query is known as the search engine selection problem.This paper proposes a new solution to the search engine selection problem. The proposed solutions borrows Summary Schemas Model from the multidatabase arena. It also utilizes user feedback and past performance to improve upon the future performance.A simulator was developed to show the validity of the model and to test its performance against solutions embedded in the design of some meta-search engines. Simulations results are presented and analyzed.