Abstract: We consider a failure-prone machine whose up and down times follows general probability distributions. A policy called preventive hedging point control, as well as a parameter optimization algorithm, is proposed. Using perturbation analysis technique, the sensitivity estimate of the performance index with respect to control parameters is derived, and it is proved that the estimate is unbiased. Then, the stochastic approximation method is applied to determine the optimal control parameters. Simulations examples show the effectiveness of the policy.