With the development of cloud services, many network fraudulent incidents have damaged the interests of transaction subjects and brought a great crisis to online service. To analyze and solve the key conditions for different participants to achieve mutual trust and win-win situations in the cloud computing, we propose a trust prediction game model composed of the supervisor, cloud service provider (CSP), and user. Firstly, we use the Bayesian formula to predict the behavior trust attribute and propose the trust service map function and weight algorithm by the principle of maximum dispersion. Secondly, based on the perspective of the user, we establish a two-party dynamic game model and discuss the existence possibility of game equilibrium. Thirdly, based on the perspective of CSP, we establish a tripartite game model, analyze several possible equilibria, and obtain the equilibrium strategy from a mixed perspective. Finally, the experimental results show that the supervision has a significant impact on the strategy selection of the players, verify the correctness and effectiveness of various game parameters, exhibition the better advantage than several articles under some different standards, and make a useful exploration for building a secure and trusted cloud service environment.

随着云服务的发展，许多网络欺诈事件损害了交易主体的利益，给在线服务带来了巨大的危机。为了分析和解决不同参与者在云计算中实现相互信任和双赢局面的关键条件，我们提出了由主管，云服务提供商（CSP）和用户组成的信任预测博弈模型。首先，利用贝叶斯公式对行为信任属性进行预测，并根据最大分散原理提出信任服务映射函数和权重算法。其次，从用户角度出发，建立了两方动态博弈模型，讨论了博弈均衡存在的可能性。第三，基于CSP的观点，我们建立了一个三方博弈模型，分析几种可能的平衡，并从混合的角度获得平衡策略。最后，实验结果表明，监督对玩家的策略选择有重要影响，验证各种游戏参数的正确性和有效性，在不同的标准下展示出比几篇文章更好的优势，并为建立游戏提供了有益的探索。安全可靠的云服务环境。