Objective: This paper aims to build a prescription information database

mining the rules of drugs in the prescription

and study the compatibility law. Method: Using information extraction based on regular expressions and vs 2008 programming to realize automatic extraction

and realized association rules mining in SQL Sever 2005. Result: Generated a structured prescription information database

and initial extracted prescription information 81 305.When the support degree is 10% and the confidence level is 70%

9 kinds of high frequency drus and 4 rules were geted. Conclution: The method of information extraction based on regular expressions make it be possible that make a full-scale study of prescription compatibility law. The results of data mining and the theory of traditional Chinese medicine and clinical experience were matching overall

which lay a foundation for further study.