Abstract:Starting from the offshore wind turbine fault diagnosis and condition monitoring technology,statistics of the faults are collected,the main parameters of the state monitoring are given. The temperature parameter is selected as the object of monitoring. Through the temperature timing diagram analysis,temperature monitoring is divided into three states of normal operation,fault formation and upcoming failure. Neural network and genetic algorithm are used for fault diagnosis. The diagnostic results showthat wind turbine upcoming failure can be predicted through temperature changes,which helps to conduct timely warning and preventive measures.