Meteorology Grid Computing aims to provide scientist with seamless, reliable, secure and inexpensive access to meteorological resources. However, the vision of achieving these goals has been encountered with many challenges such as the degree of easy-to-use and the capability of seamless automation. Fortunately, the emerging semantic web service and semantic grid technologies will bring some significant solutions for the above challenges. Currently, Semantic Meteorology Grid Architecture (SMGA) is being developed by us to support automated meteorology grid service discovery, selection, and composition. In this paper, we present a model for semantic annotation and publication of meteorology grid services by combining the expressive power of the present ontology language and the advantages of semantic web techniques. The presented model is an initial work to support automating the discovery, selection, and workflow composition of semantically described heterogeneous meteorological services, which offers the possibility of facilitating geographically distributed meteorological scientists to resolve complex scientific problems cooperately.