Cloud computing is a very popular information technology, to obtain mutual trust and win-win cooperation between the users and the cloud service provider (CSP), we propose a game model-based trust prediction. Different participants have different motives and interests in the service game process. First, we use the Bayesian formula to predict the trust level of behavior attribute, further propose trust decision function and weight algorithm of trust attribute to ensure the accuracy by maximum dispersion principle. Second, we discuss the existence possibility of the game equilibrium and establishes a dynamic game model from the short-term benefit of the user. Third, from the long-term benefit of the CSP, we establish a tripartite game model that has multiple possible equilibria that can be converted into a mixed strategy under specific common conditions. Finally, we design simulation experiments to verify the regulation and several related factors have an important influence to improve trust and honesty and control the malicious behavior of participants; further, we compare and discuss the research results with other papers in the real data set scenario, and show the advantages and highlight value.

云计算是一种非常流行的信息技术，为了获得用户与云服务提供商（CSP）之间的相互信任和双赢合作，我们提出了一种基于游戏模型的信任预测。在服务游戏过程中，不同的参与者有不同的动机和兴趣。首先，利用贝叶斯公式对行为属性的信任度进行预测，进一步提出信任决策函数和信任属性权重算法，以最大分散原则保证准确性。其次，我们讨论了博弈均衡的存在可能性，并从用户的短期利益出发建立了动态​​博弈模型。第三，从CSP的长期利益来看，我们建立了一个三方博弈模型，该模型具有多种可能的平衡，可以在特定的常见条件下将其转换为混合策略。最后，我们设计了仿真实验，以验证规则和几个相关因素对提高信任和诚实度以及控制参与者的恶意行为具有重要影响。此外，我们在实际数据集场景中与其他论文进行了比较和讨论，并显示了优势和突出价值。