Although the intelligent design system is widely used in modern mechanical products design, the optimization of mechanical products design parameters and the method of knowledge acquisition etc. remain the key factors that restrict the development of intelligent design system. On the basis of analyzing data mining methods and knowledge based engineering (KBE) principles, an information model was built to find out the minimum attribute set of mechanical products design parameters with the improved ID3 algorithm. Then a large numerous of finite element simulation results, combined with the rough set theory and data mining technology, was used to extract knowledge and rules and to fulfill the knowledge expression and reuse. Finally a gantry crane intelligent design system was established by adopting automation technology with the secondary development of CATIA and integrating the KBE extended module with CATIA. The example shows that the proposed gantry crane intelligent design system makes up for the deficiencies of the traditional intelligent design system and improves the design efficiency and the ability to respond to market.