This thesis addresses a new behavioral model of customers which aims at grounding both actual purchasing and pedestrian behaviors in a supermarket. For this purpose, firstly we observe customer’s flow line with Radio Frequency Identification and analyze the Point of Sales data together with those obtained from the field survey. The main findings are that we can classify customers by flow line and that migratory behavior has redundancy. Based on them, we develop a model and pursue computational experiments. As a result, we confirm that reproducibility has improved to some extent.