<p>The double-layer structure of clad Al alloys led to the difficulty in predicting fatigue life. This study focused on 7075-T6 aero clad Al alloy. A hybrid fatigue life data-physical model-driven approach is proposed to predict fatigue life. Initially, the fatigue life of the 7075-T6 clad Al alloy was evaluated. Fatigue failure mechanism of the Al alloy was revealed. Then, the fatigue crack initiation and propagation life were established based on dislocation theory and fracture mechanics. In addition, the LightGBM model was trained on the fatigue life data, with its hyperparameters optimized using the gray wolf optimization (GWO) algorithm. The GWO-LightGBM model was optimized to significantly improve its prediction performance. The evaluation metrics show coefficient of determination (<i>R</i>²) of 0.950, mean absolute percentage error of 10.27% and percent bias (PBIAS) of 2.21%. Subsequently, the key parameters in the fatigue life physical model were determined using the L-BFGS-B algorithm by minimizing the difference between the physical model outputs and the GWO-LightGBM predictions. Consequently, the fatigue life predication equation of the 7075-T6 clad Al alloy was found. The method of establishing physical equation for fatigue life prediction proposed in this study provides new insight for fatigue life prediction of cladding metallic materials.</p>