On the analysis about the problem of information retrieval in the Electronic Commerce environment, this paper presents a Bayesian network retrieval model. This model adopts the topology of three layer nodes, and uses co-occurrence analysis method to mine relationships between the terms. A query expansion method according to domain ontology is used to extend the users query. Finally the similarity between the document and the query can be measured by calculating the posterior probability of relevance of the document. Experiments show that the model which will provide a theoretical basis for the problem of information retrieval in the Electronic Commerce environment can effectively improve the retrieval performance.