Cloud computing has drawn much attention in industry due to its cost-efficient schema along with more prospects, such as elasticity and scalability nature of cloud computing. One of the main service models of a cloud is software as a service, where many web services are published and hosted in the cloud. Many web services offered in a cloud have the same functional requirements, but are different in non-functional requirements such as Quality of Service (QoS). In addition, as individual web services are limited in their capability, there is a need for composing existing services to create new functionality in the form of composite service to fulfill the requirements of cloud user certain processes. In this thesis, a new model is proposed to compose web services, which is based on an agent that acts as mediator between user and cloud computing, it includes the discovery module, selection module, fuzzy logic module, execution module and adaptation module. Then, the agent will employ the fuzzy rules to select the best composition plan that meets the user requirements. It will also identify the Quality of Service criteria (price, duration, availability, reputation and reliability), which will be used in this thesis. Finally, this simulation is done in the open source environment CloudSim in cloud computing. Keywords: Web service, QoS, Web service composition, Fuzzy rules.