It is essential to consider the proper position and volume of riser during the design of feeding system. In this paper, we are presenting a new method of designing optimal riser in the steel casting processes. A technique of slitting arbitrarily 3D entity model by STL (Stereolithography) was used to obtain accurate values of the partial modulus of casting, and then a mathematical model of the process of the riser design was optimized by a genetic algorithm (GA); with the help of the CAE system, which has an ability to calculate automatically and verify the validity of the optimized results, we will pursue the goal of obtaining the desired riser with optimal size and distribution but without causing any defect in casting. Thus, by combining the numerical optimization with the traditional riser design, our method proposed here will be more practical and reliable. In the end, we give an example to demonstrate the feasibility of our optimized approach in the riser design. Keywords:Riser design, Genetic Algorithm, casting optimization, STL