Developing process variants enables enterprises to effectively adapt their business models to different markets. Existing approaches focus on business process models to support the variant development. The assignment of services in a business process, which ensures the process variability, has not been widely examined. In this paper, we present an innovative approach that focuses on component services instead of process models. We target to recommend services to a selected position in a business process. We define the service composition context as the relationships between a service and its neighbors. We compute the similarity between services based on the matching of their composition contexts. Then, we propose a query language that considers the composition context matching for service querying. We developed an application to demonstrate our approach and performed different experiments on a public dataset of real process models. Experimental results show that our approach is feasible and efficient.