The fault vibration signal of planetary bearing is affected by complex transm ission paths,strong background noise andgear vibration interference,which makes the fault features weak and difficult to extract.To address these issues,a parameter adaptive multi point optimal minimum entropy deconvolution adjusted(PA-MOMEDA)is proposed to extract the weak fault featuresof planetary bearing.In order to overcome the shortcomings of MOMEDA relying on human experience to select the main parameters,a new multirobjective optimization index is established,and the optimal parameters of MOMEDA are automatically determined by the particle swarms optimization algorithm.The MOMEDA with the optimal parameters is utilized to deconvolve theplanetary bearing fault signal.Aiming at the problem of the serious edge effect of MOMEDA,a waveform extension strateg is designed to adaptively compensate the unconvoluted signal,which significantly enhances the deconvolution ability of MOMEDA forweak fault features.Envelope demodulation processing for the enhanced deconvolution signal is caried out to extract fault characteristic frequencies and identify fault type.The feasibility of the proposed method is validated using both the simulated planetarybearing signal and practical experimental signals.Moreover,compared with the traditional MOMED,MCKD and fast spectral kurtosis methods,the proposed method can extract weak fault impact characteristics and realize the accurate diagnosis of the planetarybearing fault.