To address the problem of non-consistency and overlapping influence of multi-fault impact resonance bands of rolling bearings, a synchronous diagnosis method for multi-fault bearings based on the cyclic pulse index (CPI) spectrogram is proposed. Firstly, the variation coefficient of the short-time pulse peak moment is taken as the cyclic pulse index to quantitatively characterize the cyclic periodicity and impulsiveness of bearing fault impacts. Then, by combining the frequency band tower decomposition of adaptive redundant lifting wavelet packet with the CPI calculation for the individual frequency band signals, the CPI ratio spectrogram (CPIRgram) is constructed. The optimal resonance band of bearing fault signal is adaptively selected according to the principle of the maximum CPI ratio. Finally, the cyclic pulse spectrum is employed to uniformly characterize each fault feature frequency of the bearing. The simulation and fault test results show that this method does not require prior knowledge of faults or optimization of decomposition parameters. It can accurately detect multiple fault feature frequencies even in the presence of strong noise and random transient interference. The detected fault frequencies show an error of less than 1. 6 Hz compared to their theoretical values. Additionally, the method demonstrates good robustness to variations in fault impact intensity and impact mode, indicating strong potential for practical applications.