According to the design requirements of heat sink applied in high-power LED for increasing cooling performance but reducing mass, a two-objective optimization design, including junction temperature and fins mass, is performed respectively based on orthogonal experiment design method and genetic algorithm. In early stages of optimization design, some structural variables for optimization such as fin spacing, fin height and fin thickness are determined through referring to previous researching results. And the constraint of each structural variable is assigned according to the suggestion structure. On this basis, the junction temperatures and fins masses are calculated respectively through thermal circuit method and calculation formula of fin mass, for the different combination cases of above three structural variables. In the process of optimization design, the implementation schemes of above tow optimization methods are illustrated in detail. And both of these optimization methods need to construct the evaluation function of two-objective optimization by applying a weighted sum method. The comparative results between two types of objective values before and after optimization show that, both of optimization methods can offer global optimal solution on Pareto front, and the optimization effects are all obvious, in addition, the genetic algorithm can provide more diversified optimized structures.