对机器故障为一般概率分布的不可靠制造系统，提出一种预防阈值控制策略和实现这一控制策略的参数优化方法．应用扰动分析法得到性能指标对控制参数的灵敏度估计，并证明了估计的无偏性，又利用随机逼近技术优化控制参数．仿真结果验证了控制策略的有效性 For an unreliable manufacturing system whose machine fault is a general probability distribution, a threshold control strategy and parameter optimization method to realize this control strategy are proposed. By using the perturbation analysis method, the sensitivity of the performance index to the control parameter is estimated, and the unbiasedness of the estimation is proved. Moreover, the stochastic approximation technique is used to optimize the control parameters. Simulation results verify the effectiveness of the control strategy