In the long term use process,the performance of roling bearing willbe degraded to different degrees.If the degradationstate of roling bearing can be identified online,accidents can be effectively prevented.In this paper,an adaptive noiseassisted collective empirical mode decomposition(CEEMDAN)method combined with energy entropy is proposed to extract the characteristics of vibration signals,and then the characteristics are input into the DSHDD model,and the obtained results are input into themembership function to calculate the membership,which can be used as the evaluation index of performance degradation.An adaptive threshold is set using 3o to determine the bearing's early failure threshold.CEEMDAN and Hibert envelope demodulationmethods are used to verify the correctness of the evaluation results.The validity and practicability of the model are verifed by usingthe bearing life cycle data from the University of Cincinnati.