The search for structural fragments (substructures) of compounds is very important in medicinal chemistry, QSAR, spectroscopy, and many other fields. In the last decade, with the development of hardware and evolution of database technologies, more and more chemical compound database applications have been developed along with interfaces of searching for targets based on user input. Due to the algorithmic complexity of structure comparison, essentially a graph isomorphism problem, the current applications mainly work by the approximation of the comparison problem based on certain chemical perceptions and their search interfaces are often e-mail based. The procedure of approximation usually invokes subjective assumption. Therefore, the accuracy of the search is undermined, which may not be acceptable for researchers because in a time-consuming drug design, accuracy is always the first priority. In this dissertation, a design of a search engine for chemical compound database is presented.The design focuses on providing a solution to develop an accurate and fast search engine without sacrificing performance. The solution is comprehensive in a way that a series of related problems were addressed throughout the dissertation with proposed methods. Based on the design, a flexible computing model working for compound search engine can be established and the model can be easily applied to other applications as well. To verify the solution in a practical manner, an implementation based on the presented solution was developed. The implementation clarifies the coupling between theoretic design and technique development. In addition, a workable implementation can be deployed to test the efficiency and effectiveness of the design under variant of experimental data.