We in this paper explore a new mining paradigm, called Indoor Traversal Patterns (abbreviated as ITP), to discover user traversal behavior in the mall-like indoor environment. The ITP algorithm can identify user traversal sequences from uncertain user itineraries with the RFID-based indoor positioning technology. Note that it is a highly challenging issue in the indoor environment to retrieve the precise locations in the indoor environment. Since previous works on mining user moving patterns usually rely on the precise spatiotemporal information from GPS signals, it is difficult to apply similar approaches to discover user traversal behavior in the indoor environment. We therefore develop a framework to transform the RFID antenna data to uncertain user traversal transactions, and further diminish the uncertainty before mining the indoor traversal patterns. Our experimental studies show that the proposed ITP algorithm can effectively overcome the impact from location uncertainty and discover high-quality traversal patterns, to provide insightful observation for marketing decision.