The fault features are extracted based on the ensemble empirical mode decomposition and autoregressive（ EEMD-AR） model

and the deep belief network（ DBN） is used to realize high precision fault diagnosis of the bearing of the wind turbine. The bearing vibration signal is decomposed by the improved EEMD

the first four IMF components are selected to establish an AR model

and the autoregressive parameters of the AR model are obtained by the least squares. The autoregressive coefficient is used as the input parameter of DBN

and the batch normalization layer is added to optimize the DBN

and the representative high-order features are mined from the signal by DBN to realize the fault classification of the bearing. The simulation experiment shows that the method can extract the fault characteristics of different faults of the wind turbine bearing and has higher accuracy than the traditional BPNN and SVM.