Under market surroundings, the demands of market to products are dynamic. The primary static production pattern is difficult to suit it well because of larger lag time. To change the situation and optimize resources allocation, based on singular and lattice order rough set of their dynamic characteristics and knowledge discovery abilities, in this paper we present a novel modeling method for products dispatch procedure under dynamic environment. Some unexpected events such as urgent processing and market demands alteration are also fully considered, reselection and rescheduling of the products inside real-time window are also studied. Rough model of real-time dispatch for the products is established based on single rough set. Then the products within the scope of real-time window are ranked in light of their manufacture priorities based on lattice rough set. In the end, a simulation trial indicates the proposed method can better adapt to dynamic indeterminate environment and reduce the scale of solution, and is an effective real-time dispatch strategy.