This thesis focuses on the problem of Web services in distributed and heterogeneous systems. More specifically, it will be to focus on a particular type of services called Data Web Services (DaaS: Data as a Service). The objective of this work is to propose extensions DaaS compositions methods compositions in an environment characterized by a strong presence of uncertainty. Indeed, in a context such as the Internet, it is increasingly recognized that the data and therefore the data access services are subject to uncertainty values while requiring more sophisticated management techniques. In this thesis, it will matter to enrich the semantic annotations of Web services to reflect this dimension of uncertainty, and then propose appropriate compositions mechanisms. Such mechanisms can be based on probabilistic or possibilistic theories to consider different possible worlds’ interpretation of uncertain data.