Learning Grid services are the fundamental component of learning systems based on Grid Technology and represent functionalities that can be easily reused without knowing the details of how services have been implemented. Semantic Web Service technology promises to automate web service discovery, composition and integration, tasks that currently need to be performed manually despite the quickly increasing number of online service. A number of approaches have been proposed to tackle the problems of semantic description and matchmaking of Learning Grid Services. This article gives an overview of those recent research efforts.