The 1.5-dimensional spectrum is widely used in the field of fault diagnosis for its excellent feature against with Gaussian white noise. The 1.5-dimensional energy spectrum combined with energy operator demodulation and 1.5-dimensional spectrum is more effective for bearing fault diagnosis. However, this method has a poor effect in dealing with low signal-to-noise-ratio signal. In order to solve the problem of fault feature extraction under strong background noise, a diagnosis method combining minimum entropy deconvolution（MED）with the 1.5-dimensional energy spectrum was proposed. The original vibration signal was first denoised by using MED, and then the processed signal was processed by using 1.5-dimensional energy spectrum; the frequency components in the envelope spectrum was analyzed and compared with the corresponding fault feature frequency to obtain the diagnosis result. The effectiveness and superiority of the proposed method were verified by simulation data and a variety of measured data.