Ontology is a formal tool to model sharing the knowledge of the semantic Web, which is based on the description logic. Dynamic description logic handles static and dynamic knowledge in a uniform frame. There is much of dynamic knowledge with uncertainties in the semantic Web. Against this background, this paper proposes an extended knowledge representation method called Cloudbased Dynamic Description Logic (CDDL for short), i.e., a variant of DDL based on the Cloud model. CDDL has the ability to represent and reason various kinds of knowledge efficiently, including the static and dynamic knowledge with uncertainty. Compared with some traditional description logic, CDDL can express dynamic knowledge, which is necessary in the semantic Web. CDDL can deal with uncertainties more extensively, and model the knowledge in the semantic Web more reasonably. Moreover, some examples and analysis illustrate that the Cloudbased method is feasible and effective, and in a certain sense, CDDL is more suitable to be used as the basis of logic in the semantic Web.