Abstract: Aiming at the automatic service composition of remotely sensed information processing service,this paper proposes a semantic matching based composition method.Based on remote sensing information ontology,a hierarchical semantic matching methods on both service category and remotely sensed image metadata is introduced.And a progressive composition is implemented by two steps:① the service and/or graph related to user requirement is created on-line to limit the services' selection range;② AO* based heuristic search algorithm is used for optimal path in the and/or graph.Typical remote sensing change detection application is employed as a case study.