In order to monitor the agglomeration in gas-solid fluidized bed, the voiceprint characteristics should be accurately extracted. Acoustic signals created by the polymer impact on the inner bed wall of the fluidized bed reactor were monitored with pasting piezoelectric ceramic sensors on the outside wall of fluidized bed reactor, transmitting the charge with long shielded cables and audio sampling. The original signal collected during the period of the fluidized bed working condition from normal to agglomeration was decomposed into seven sub-signals. When the particle size in the fluidized bed increases, the energy percentage of d6, d7 wavelet scale will rise. The energy percentage of the d6, d7 wavelet scale can be the representative of the particle size change in the fluidized bed, which can be used as the voiceprint characteristics to monitor the fluidized bed agglomeration fault.