At present, the existing descriptions of trust by trust valuation models lack comprehensiveness. To solve this problem, this paper discusses the co-existence and integration of fuzziness and randomness of trust relation, analyzes the ways cloud models describe uncertain concepts and the algorithms cloud models transform between qualitative concepts and their quantitative expressions, and presents trust cloud, a trust evaluation model based on cloud theory. This model provides the algorithms of trust information transfer and combination described by digital characteristics. While describing accurate trust expectation, this model portrays the uncertainty through entropy and hyper entropy. Compared with traditional trust valuation models, trust values obtained in this model contains more semantic information, indicating that valuation results from the model are more suitable as evidence for decision-making on trust.