Cyberinfrastructure aims to cooperate information and communication technologies to deal with growing volume and heterogeneity of data, information, and knowledge. In GIScience, spatial data infrastructure, cloud computing, and Web services are some research issues that aim for geospatial information cyberinfrastructure. Although these research issues have paved a way for data interoperability in the cyberinfrastructure, how to cooperate with knowledge is still a challenge (e.g. knowledge sharing and reusing, and knowledge-oriented discovering needed data). Therefore, in this research, we provided a knowledge-oriented framework for cyberinfrastructure to integrate knowledge from domain experts, data from data providers, and technologies by using ontologies in the four parts of the framework: a Web portal, an ontologies engine, an ontology registry in SDI, and a service chain mediator. The Web portal is for users (e.g. decision makers) to submit their geospatial statements. By ontologies which manage geospatial knowledge, contributed by domain experts, in a standard, structural, and sharable format, the ontologies engine can understand the structural knowledge to infer solutions in response to the submitted statement, and to automatically discover needed GI services, including data and functions, in the SDI. The chain mediator then constructs a workflow to and consequently processing the GI services to generate a result for the statement. The proposed framework presents a new flexible architecture to integrate knowledge form domain experts, multiple GI services in SDI, and technologies from ontologies for the knowledge-oriented cyberinfrastructure.