Aiming at the problem of identifying the degradation state of bearing components in rotating equipment of nuclear power plant (NPP), a method for identifying the decline state based on GG (Gath-Geva) fuzzy clustering was proposed. First, from the feature vector set consisting of the time-frequency domain of the bearing's full-life vibration signals, the spectral energy and RMS (Root Mean Square) which can better represent the trend of bearing degradation evolution were selected to form a two-dimensional feature vector. The GG clustering method was used to identify different degradation stages of bearings. The IMS (Intelligent Maintenance System) center bearing experimental data is used to analyze the case, and compared with the FCM (Fuzzy C-Means) and GK (Gustafson-Kessel) clustering methods. The verification shows that GG clustering has a higher classification performance for the degraded state of bearing components.