In industrial areas, understanding the preference of customers is one of the important considerations for establishing profitable product manufacturing plans. As one of the approaches in pattern mining, high utility pattern mining has been employed to find a set of products creating high profits by considering the purchase quantity and price of each product. In this regard, high utility pattern mining can be useful to establish profitable product manufacturing plans that allow a corporation to maximize its revenue. For establishing manufacturing plans, we also need to understand the recent preference of customers from stream data, which are continually generated without limitations. In this paper, we propose a novel algorithm and list structure for finding high utility patterns over data streams on the basis of a sliding window mode. Unlike existing algorithms, the proposed algorithm does not consume huge computational resources for verifying candidate patterns because it can avoid the generation of candidate patterns. Therefore, the algorithm efficiently works in complex dynamic systems. Experimental results obtained from various tests using real-world dataset show that the proposed algorithm outperforms state-of-the-art methods in terms of runtime, memory usage, and scalability.