Web service composition is a key technology for creating value-added services by integrating available services. With the rapid development of Service Computing, Cloud Computing, Big Data and Internet of Things, a fast increasing number of services with similar functionalities but different Quality of Service (QoS) are available on the Internet, and make Web service optimal composition a NP-hard problem. Meanwhile, over the last decade's development and evolution of the service industries, service domain features (such as priori and similarity of services) are gradually formed. These features have valuable domain knowledge for improving Web service optimal composition. However, existing research works on Web service composition don't make full use of the service domain features, and leading to unfavorable results. Therefore, how to improve the efficiency and effectiveness of Web service optimal composition with the knowledge of service domain features becomes a significant challenge. To attack this issue, this paper firstly analyzes the influences of service domain features on Web service optimal composition; then, improves key optimization strategies of ABC with the knowledge of service domain features; finally, proposes the Web service optimal composition method based on improved artificial bee colony algorithm (S-ABCSC). The performance of S-ABCSC is verified through simulation experiment; Moreover, the underlying dependencies between service domain features and S-ABCSC's optimality are analyzed and S-ABCSC's parameter settings are determined through several experiments with different service usage data sets.