In the current Web, e-document has been the most common vehicle for delivering and exchanging information. As the amount of e-documents has grown enormously, effective classification facilities are urgently needed to classify and query e-documents users want. In this paper, we propose a method to classify e-documents into a set of predefined categories based on Fuzzy Cognitive Map (FCM). The e-documents are collected from Internet by a metasearch engine. FCM has been employed to capture the semantic relationships between keywords of e-documents. Experiments with a set of local e-documents have proved that this approach has high performance and can help users getting the e-documents efficiently and effectively. The proposed method has been implemented and integrated into the Dunhuang Feitian System to manage and classify e-documents.