To overcome the difficulty of swash-plate bearing early fault features of the helicopter easily immerged in strong background noise

a method based on the minimum entropy deconvolution was proposed through the optimization design of the inverse filter

its kurtosis value was increased by about 3.8 times

the vibration signal was denoised and the weak fault characteristics of vibration signal were enhanced. By using local mean decomposition (LMD) and kurtosis-correlation coefficient criterion to settle the issue of non-stationary and nonlinear signal frequency components of complex demodulation

it can adaptively decompose and optimally reconstruct the non-stationary and nonlinear signal

and improve the signal-to-noise ratio of vibration signal. To settle the issue of coupling modulation and sideband highlighting

through noise reduction in envelope spectrum by introducing the general Shannon entropy

the amplitude ratio of the first-order fault characteristic modulation frequency to the fault characteristic frequency was reduced by about 24%-43%. The diagnosis research results of laboratory signal and the fault diagnosis test system for a type of helicopter swash plate bearing verify that the method is reliable and applicable.