中國古籍數碼化工作開展多年，建設了諸多成果卓著和類型豐富的古籍數據庫。這些數據庫數量龐大，在內容、技術框架、功能設置等方面存在著諸多共性和差異，亟待從其形態和功能上做出劃分和論證。文章在調研了國內外280個古籍數據庫的基礎上，將其分為四個層次，即古籍影像庫、古籍全文庫、古籍語料庫和古籍知識庫。每種類型古籍數據庫都有著其不可替代的功能和作用，使用者可以根據自身需求做出不同的選擇。同時，數據庫建設機構可以更好地以知識庫為建設重點，開展有針對性的建設規劃。

The digitization of Chinese ancient books has been carried out for decades, with a large number of different types of databases. Each database has its own contents, characteristics, technical frames, and functions, which has to be focused to divide their types and functions. Based on the investigation of 280 ancient book databases in the world, this paper divides them into four layers, namely image database, full-text database, corpus and knowledge database. All types of ancient book databases have their own irreplaceable role, depending on the needs from the users.The constructors of the databases can pay more attentions on the knowledge base in their plan for ancient book database construction.