Due to the exponential growth in demand from e-commerce businesses, warehouses are increasing in size to accommodate dynamic needs. Among the various activities in warehouses, order picking accounts for approximately 55% of the total operational costs. Mobile robots capable of lifting shelves to picking stations are employed to perform the order picking operation in these warehouses to meet rising demands. In a warehouse with multiple order-picking robots and dynamically incoming orders, assigning specific orders to each robot is a highly complex task. Optimal task allocation can enhance warehouse efficiency, thereby reducing operational costs. In the present work, a clustering algorithm is used for optimal task allocation to the order-picking robots, based on the distance between the robots, the picking station, and the shelves to be picked. Simulation experiments were conducted by varying the number of robots, as well as their positions relative to the picking station and the shelves to be picked. The evaluation covered increasingly large warehouse scenarios, ranging from 8 robots handling 16 tasks to 18 robots handling 40 tasks, enabling performance comparison of the allocation strategies under higher task–robot loads. The result indicates that the iterative K-means clustering successfully reduces task distances, enhancing overall warehouse efficiency when compared to established task allocation approaches.