In Web data integration, selecting data from a Web data source collection such that the specific query intents are satisfied while the number of accesses to data sources is minimized and the quality of returned results are guaranteed is a popular topic. In this paper, using the researches and practices in recent ten years as the background, the study focuses on the evolution and presents research in the area of Web data source selection and classifies Web data source selection methods. In addition, the paper discusses the research motivations, methods and results of relevance-based data source selection and quality-based data source selection. Moreover, the paper introduces the correlation research results and analyzes their destinations, key techniques, merits and demerits. Finally, some directions for future research are put forward.