Inspired by service computing principles, in cloud manufacturing, manufacturers encapsulate their resources into consumable services that can be looked up and accessed over the Internet. Manufacturing ontologies are used to store the service information. Manufacturers use service rules to control how their resources can be accessed. The rules are normally written in natural language. Thus, they need to be converted to semantic rules that can be understood by the search engine of the manufacturing ontologies. Manually converting service rules to semantic rules is time-consuming and error-prone. This paper proposed an approach that automatically converts service rules to semantic rules. The proposed scheme classifies the semantics of typical service rules into several semantic categories. Natural language processing techniques are used to process the service rules to map the semantic meanings of the rules to the relevant semantic categories. Then, the identified semantic categories are converted to semantic rules. The evaluation of the scheme shows that the scheme achieves good conversion accuracy.