Scientific researchers usually need to deal with a large amount of scientific and technical articles. A handy toolkit to facilitate retrieval and identification of these articles is vital. This paper presents an Agent-based Personal Article Citation Assistant (APACA). As an autonomous agent, it provides extensive and efficient assistance for users to identify most related citations in a local article repository in collaboration and interaction with other distributed personal citation assistants. Furthermore, this paper proposes an Article Ontology and an optimised Bayesian network method to infer the most relevant article annotation in the local Article Ontology using a given keywords set K.