For the difficulty of predicting the life of gear,an unbiased Grey GM（1,1）model is established based on the original data,and it is modified by using residual error model.The development trend of gear life is analyzed,the predicting accuracy is improved by introducing Markov optimization based on unbiased Grey GM（1,1）modification model.Comparing with the predicted value and the actual value,the predicting accuracy of gear life is improved from 87.06%to 90.09%.The result shows that the combined model has a better predicting precision and practicality within a period in the case of few data and poor information,it provides an available method for predicting the life of gear.