中醫古籍文獻整理工作，交由電腦代勞是相當經濟實用的方法。中文全文自動擷取技術已相當成熟，應用在中醫古籍照樣可行。本研究選取電子河間六書為測試專案，河間六書全文資料庫共190431 字，本草備要藥物名詞庫共選取203 種藥物，專屬河間六書藥物名詞庫共選取229 種藥物（含前項 203 種藥物）。電腦程序運用上述之藥物資料庫擷取全文有關方劑專論之藥物組成，結果篩選出603 組藥物群組，其精確率82% ，回收率為95% ，對輔助人工整理中醫藥古籍資料效率相當高。研究最後由人工方式核對資料，共整理出河間方劑512 首，並已登錄在中醫藥古籍文獻網際網路上供使用。本研究的電腦程序、擷取流程及藥物名詞資料庫可移植到任何中醫藥古籍文獻資料上重覆使用。Information can be retrieved from Traditional Chinese Medicinal Literature (TCML) with the aid of the computer and considerable savings in time and money. The Chinese full text information retrieval technology is well developed and already applied to the TCML. In this study, the electronic text of Heijian-Liu-Shu (The Six Books of Hejian) was chosen as a pilot project. This book in form of database contents totally 190431 Chinese characters, the computer program matched 203 medicinals in the Ben-Cao-Bei-Yao , and finally 229 medicinals were filtered from the Heijian-Liu-Shu (including the above 203). The filtered medicinals database was used to extract ingredients from the Heijian-Liu-Shu formulas automatically. The results show that 603 ingredient groupings were indexed and the precision ratio was 82% and the recall ratio 95%. The efficiency gave us confidence in the area of computer-aided TCML information retrieval. Finally, quality control of 512 formulas was completed. These recipes have already been installed on the TCMET world-wide-web. All the computer algorithm, flows, procedures and the medicinals database which developed in this paper can be transposed to any other similar TCML information retrieval project flexibly.