To effectively address dynamic target variations and emergencies in battlefield environments, a DWTA (dynamic weapon target assignment) method based on a real-time rolling mechanism was proposed. An allocation model based on the cumulative probability of destruction and a weapons scheduling model aiming at the shortest operational completion time were established. To adapt to the time-varying availability of weapons, a VLC-DPSO (variable length coding discrete particle swarm optimization) algorithm was introduced, utilizing discrete optimization through copying optimal solution segments. Moreover, a hierarchical closed-loop optimization of assignment and scheduling was achieved by integrating the genetic algorithm for weapon scheduling sequence optimization. Simulation experiments show that, compared with the traditional genetic algorithm and particle swarm optimization algorithm, the proposed VLC-DPSO algorithm achieves better strike effects in both small-scale and large-scale scenarios, especially showing significant advantages in large-scale scenarios. Furthermore, experimental scenarios with randomly introduced new targets verify the proposed method"s robust adaptability to battlefield emergencies.