Abstract: Gamma process was adopted to items’ degradation process description whereas its smooth and independent increment property which is required for degradation model. Moreover, with the problem of remaining useful life prediction without sufficient failure data, a remaining useful life prediction model based on Gamma degradation process was established using the abundant indirect condition information and few direct condition information that could reflect items’ degradation state. Particle Filters was introduced to Experience Maximization (EM) arithmetic to obtain an analytical result of likelihood function. At last, the prediction model was implemented on the planetary carrier of helicopter main gear-box and had obtained the remaining useful life on 95% confidence interval, the result show that the model is effective.