Many automatic or semi-automatic extraction techniques have been proposed for building domain ontologies in recent years but the correctness, consistency and completeness of the extracted ontologies is often either not considered or is not formally verified. The issue of detecting potential anomalies in an ontology has not to date been adequately addressed. In this paper we propose a formal technique for ontology representation and inference, based on which an automatic technique for ontology verification can be developed so as to be able to detect and identify potential anomalies in an ontology. The technique makes use of a State Controlled Coloured Petri Net (SCCPN), which is a high level net that combines a Coloured Petri Net and a State Controlled Petri Net. This work presents a formal definition of SCCPN for modeling ontologies and the mapping between them as well as formulating the ontology inference in SCCPN with specified inference mechanisms.