The damage and failure of wet clutch is the main factor restricting the reliability improvement of transmission system.In order to obtain the critical damage threshold of damage characteristics of wet clutch,the friction and wear bench test of wet clutch was carried out.Taking the change rate of surface roughness，the change rate of Fe and Cu concentration，the critical wear rate as the research objects,a damage threshold model of wet clutch based on PSO-BP neural network algorithm was established.The results show that the absolute average value of the relative error between the prediction result of the threshold model and the measured value is 5.74%,and the prediction result of the threshold model is close to the measured value.The change rate of Cu concentration is one order of magnitude larger than that of Fe concentration.The wear of friction elements is mainly the wear of the copper coating on the friction plate surface.The multi wear characteristic parameters increase nonlinearly with the increase of oil pressure and speed,and the multi wear characteristic parameters are more sensitive to the change of oil pressure.