Abstract: In order to quantitatively evaluate the damage evolution process of materials during service, identify the damage state and ensure their safe operation, this study conducted acoustic emission (AE) monitoring experiments on the corrosion fatigue crack growth (CFCG) of aluminum alloys. Hierarchical clustering was employed to extract feature parameters, followed by K-means clustering. The clustering results and scanning electron microscopy (SEM) images were used to identify different types of damage during the CFCG process. Finally, the damage evolution process of aluminum alloy of CFCG was studied by analyzing cumulative counting and cumulative energy of each type of damage. The experimental results showed that the combination of unsupervised clustering of AE parameters selected by hierarchical clustering and SEM image analysis could effectively identify damage modes such as hydrogen evolution, fracture of second phases and inclusions, microcrack formation due to plastic deformation, crack propagation, and pitting corrosion in the CFCG process of aluminum alloys.