Abstract: This paper discusses several problems about the application of natural language processing to intelligent searching and information extraction,and gives some processing methods to label and perfect the dictionary-based repository.