研究密集多回波环境下的机动多目标跟踪问题。通过对多目标联合概率数据关联方法性能特征的分析，将其归结为一类约束组合优化问题。在此基础上，利用随机神经网络求解组合优化问题的策略，采用改进的增益退火算法，提出了一种新颖的机动多目标快速自适应神经网络跟踪方法。仿真结果表明，该方法不仅具有很高的收敛速度和跟踪精度，而且计算量小，关联效果好。 Research on Multi-target Maneuver Tracking in Dense Multi-echo Environment. By analyzing the performance characteristics of the multi-objective joint probability data association method, it is reduced to a class of constrained combinatorial optimization problems. On the basis of this, the strategy of solving the combinatorial optimization problem by using the stochastic neural network is proposed. A novel GA-tracking method based on the improved GA is proposed. Simulation results show that this method not only has high convergence speed and tracking accuracy, but also has a small amount of computation and good correlation effect.