When a firm develops a new product and plans time-to marketing strategy, it makes great efforts in analyzing customers’ demand. From this point of view, finding out the customer demand curve is critical in determining the optimal supply and price level of this product. However, forecasting demand curve is very costly and have many limitations in data accessibility. As E-Commerce thrives, online auction has emerged as one of main stream market mechanism beyond the roles of traditional auctions. In our research, we pay attention to a special feature of auction mechanism. Auction forces bidders to reveal their Willingness-To-Pay (or reservation price) for the goods. Online Auction produces large volume of such data. However, to the best of our knowledge, little research was conducted to utilize such data. In this paper, an exploratory approach for obtaining customer demand curve is proposed. The idea of this approach is based on that customers’ partial demand curve could be estimated through the last bidding data of online auction participants. Therefore, the practical signification of this research is proposing an intelligent process estimating customer demand curve in a cost effective and rapid way. This research is carried out based on the agent-based modeling and estimating process via censored regression.