Hazardous weather, such as thunderstorms, is one of the primary causes of airspace congestion in summer. A scientific and accurate assessment of the dynamic capacity of the terminal area is required to promote the balance between airspace capacity and traffic flow. First, the dense inverse search (DIS) optical flow algorithm was selected for thunderstorm echo extrapolation over the next 0~2 h. Based on this prediction, restricted flight areas were delineated dynamically. Subsequently, a terminal area network was established. Control transfer points, the airport, and restricted areas were set as nodes, and flyable paths were defined as arcs. The maximum flow of the network was determined using the max-flow min-cut theorem. Finally, the capacity in various directions and the total capacity were calculated by altering the positions of source and sink points. The results indicate that the DIS optical flow algorithm effectively captures the random generation and dissipation of thunderstorms. Combined with the semi-Lagrangian extrapolation scheme, the method quantifies the impact of weather evolution on airspace capacity. The predicted flight restricted areas are more precise than those from traditional methods, enabling a rolling assessment of dynamic terminal capacity for the next 0~2 h at 6-minute intervals.