Wind turbine blade bearings are often operated in harsh circumstances which may easily be damaged causing the turbine to lose control and to further result in the reduction of energy production. However, for condition monitoring and fault diagnosis (CMFD) of wind turbine blade bearings, one of the main difficulties is that the rotation speeds of blade bearings are very slow (less than 5 rpm). Over the past few years, Acoustic Emission (AE) analysis has been used to carry out bearing CMFD. This article presents the results that reflect the potential of the AE analysis for diagnosing a slow-speed wind turbine blade bearing. To undertake this experiment, a 15-year-old naturally damaged industrial and slow-speed blade bearing is used for our study. However, due to very slow rotation speed conditions, the fault signals are very weak and masked by heavy noise disturbances. To denoise the raw AE signals, we propose a novel cepstrum editing method, discrete/random separation based cepstrum editing liftering (DRS-CEL), to extract weak fault features from raw AE signals, where DRS is used to edit the cepstrum. Thereafter, the morphological envelope analysis is employed to further filter the residual noise leaked from DRS-CEL and demodulate the denoised signal, so the specific bearing fault type can be inferred in the frequency domain. The diagnostic framework combining DRS-CEL and morphological analysis is validated by comparing several methods and related studies which offers a promising solution for wind farm applications.