摘要：Preform design for shell nosing product without machining the inner surface after forming is an experience-extensive work. Generally, the initial design needs to be modified and simulated consequently to get proper preform and nosing die, and the iterative process is time consuming. This paper puts forward a new approach, in which the suitable initial design can be obtained by knowledge-based intelligent technology and the optimal design can be acquired by finite element method （FEM） based geometrical modification. With the cax object model as the bridge of CAD and CAE, the CAD model with simulation knowledge can be transferred into CAE automatically, and the CAE result can be automatically utilized as well. Based on the comparison between the simulated shape and the desired shape, the dissatisfactory area will be modified. A new simulation and modification process will be carried out based on the modified design. The process is repeated iteratively to get the optimal design. This approach utilizes the commercial CAD and CAE software, without the need of complex back-forward FEM procedures. Based on the new approach, an in-home intelligent shell nosing design and optimization system is developed, and a case study proves that this system can reach a reasonable design efficiently.