The invention provides a potential customer behavior analysis method and a potential customer behavior analysis system based on face recognition. The potential customer behavior analysis method comprises the steps of: photographing a customer arriving at a store by means of a camera, extracting feature information and generating face data when a face image is detected, comparing the face data withexisting face data, analyzing the data according to a matching result, if the customer is a potential customer, outputting sorting information and sending an information template A to a shop assistant terminal at the same time, and if the customer is not the potential customer, sending an information template B to the shop assistant terminal; and creating a customer information table for the customer upon unsuccessful matching, and outputting and storing sequencing information. Therefore, recommendation for potential customers is realized, the coverage of recommended crowds is improved, the behavior of the user is recorded and analyzed based on face recognition instead of only performing business analysis recommendation on purchase records of existing member customers, an effective recommendation result is obtained, and meanwhile, data support is provided for developing new members of the store and improving the sales efficiency.