Abstract To face the challenge of the upcoming knowledge-intensive industries, the current state-of-the-art of mechanical CAD (Computer-Aided Design) technique, which is based on numerical calculation combined with a computer graphics function, cannot fully meet the need of decision-making automation. Any complex mechanical design generally consists of two different processes, i.e., qualitative reasoning and quantitative computation. The qualitative decision-making of design, such as conceptual design, is usually hard to depict with numerical models. It is becoming apparent that a new methodology is demanding the integration of numerical processing and symbolic reasoning, to integrate different domain knowledge as well as to integrate different objectives and functions into a complete software environment for more advanced design automation. This paper presents a parallel hierarchical structure of an integrated intelligent system and application in mechanical design.