In order to provide serious game products with player experience as its core, this paper proposes a serious game design method of learning methods game mechanism model, fuzzy Kano model and Entropy TOPSIS method. Firstly, according to the goal of game and related paradigms, a targeted learning mechanism-game mechanism model is established, and then the relevance of the indicators in the model is evaluated to screen out the design elements that initially accord with the theme of the game and the purpose of education. Secondly, the fuzzy Kano model is used to design the questionnaire survey to get the demand classification. Taking the essential demand as the core design element, the expected demand and excited demand are extracted to sort, thereby calculating the satisfaction index. Thirdly, the evaluation information of satisfaction index is collected by Likert scale, and the objective weight of design elements is obtained by entropy method, and the importance order of each design element is determined by TOPSIS. Finally, taking the entity serious game "data capital" as a design case, the feasibility and superiority of the design method are verified. In conclusion, this method can be conducive to developing serious games, improving the satisfaction of game players, and providing reference for other similar products.