As a key component of rotating machinery, the health condition of rolling bearings is closely related to the safe operation of the equipment. Timely detection of bearing faults can prevent potential safety accidents. In practical engineering, bearing fault information is frequently submerged by strong noise, making it difficult to accurately extract fault features from vibration signals. This paper proposes a fault feature enhancement method for rolling bearings based on Adaptive Ensemble Noise-Reconstruction EMD (AENEMD) and Periodic Modulation Generalized Intensity (PMGI). First, a High-order Derivable Hard Threshold Function was constructed to accurately evaluate the noise. Subsequently, the method for selecting the optimal denoising threshold of the Intrinsic Mode Function (IMF) was investigated based on Stein’s Unbiased Risk Estimate (SURE) and Gradient Descent Method. Then, the PMGI algorithm was adopted to evaluate the information richness of each IMF, and the bearing fault characteristics were enhanced through the envelope spectrum analysis of the reconstructed signal. Finally, the proposed AENEMD-PMGI was used to analyze the experimental data of the bearings, achieving fault feature enhancement for both inner and outer race defects under the − 11 dB SNR. The method proposed in this paper can provide a basis for the fault diagnosis of bearings in a strong noise environment.