The optimization of component size aimed to minimize the weight and volume is an important issue to a nuclear power plant design. The optimization algorithm is applied to determine the optimum structural and operation parameters of components in order to reduce the difficulty of manufacture and transport of large components in those nuclear power plants with high powers. A preliminary optimal design is performed on the moisture separator reheater of Qinshan I nuclear power plant in this paper. A self-adapting multi-complex-shape algorithm has been developed and coupled with the thermal mathematical model of the moisture separator reheater, written in C#. The steady-state thermal mathematical model of moisture separator reheater is used to estimate the weight of moisture separator reheater according to the variation of a given number of structural and operation parameters. Then, the thermal mathematical model coupled with the self-adapting multi-complex-shape algorithm permits a set of best structural and operation parameters of the moisture separator reheater to be determined. The results clearly show that there is plenty of potential to improve the design of moisture separator reheater, while satisfying the structural and performance constraints. The results can also help both designers and operators to make a more convenient selection of the structural and operation parameters of moisture separator reheater.