Customer behavior in a store was predicted by using Cellular Automata. Customers move around in a store depending on the layout of items, so the layouts of a store have large effect on amount of sales. But in fact, it is difficult for retailers to design the best layout. In a store, there are large amount of items which influence customers. Customer behavior can be considered as one of Complex Systems. Therefore we modeled it by using Cellular Automata (CA). CA is an useful modeling method for Complex Systems. In CA algorithm, local neighbor rules are defined as interaction of elements which compose the phenomena. By using CA, we modeled the relationships between customers and items. In the simulation, purchase history was also considered by using past purchase data which are called POS(Point of sales) data. As a result, the customer movement depending on the layout of items was predicted.