提出利用神经网络和模拟退火技术来求解有约束的FMS资源调度问题的一种新方法。有约束的FMS资源调度被分解为一系列时间间隔的调度，这些时间间隔的调度由事件驱动，随着这些时间间隔的调度的完成，整个调度过程结束。仿真结果表明，这种方法能以较快的求解速度得到全局最优解。 A new method to solve constrained FMS resource scheduling problem using neural network and simulated annealing is proposed. The constrained FMS resource schedule is decomposed into a series of time-spaced schedules that are event-driven and schedule as the completion of the scheduling of these time intervals. The simulation results show that this method can get the global optimal solution with faster solution speed.