In engineering technology, it is necessary to compare different designs in an optimization project. Therefore a big problem is how to obtain finite element analysis (FEA) data involving different parameters and how to acquire optimization parameters from the results of FEA to solve the optimization problem rapidly. The parametric mesh (Paramesh) method, automatically based on history of complicated special surfaces, is developed to obtain many results of FEA involving different parameters. This paper is presented to demonstrate the method of parametric finite element analysis (PFEA). The optimization method of process parameters optimization is based on combining PFEA/ ANN (artificial nerve net)/GA (genetic arithmetic) to find out optimization parameters. This allows one to rapidly obtain optimization parameters during a design by doing FEA only once. The research indicates that the parameters optimization method based on PFEA/ANN/GA in the product design can short the product development cycle, decrease material consumption and guarantee product quality, etc.