A novel multimodal data fusion model based on a three-dimensional convolutional neural network (3D-CNN) has been developed to predict the fatigue life of laser powder bed fusion (LPBF) additively manufactured Ti-6Al-4V alloy, as reported by Jianrui Zhang et al. in the International Journal of Fatigue. The model concatenates three-dimensional defect voxels in the region of interest with one-dimensional modal data—including process parameters, mechanical properties, and loading conditions—to achieve cross-modal synergistic modeling of text and image attributes, enabling end-to-end fatigue life prediction. To validate performance, the research team also constructed a defect-feature-based deep neural network as a baseline, and results showed that the multimodal model significantly outperformed in prediction accuracy. Fatigue performance of additively manufactured components is influenced by multiple interdependent factors, making traditional methods inadequate for precise prediction. Leveraging expert knowledge, the study defined nine defect feature descriptors and employed the SHAP method to quantify the impact of different defect features on fatigue life. This work establishes a scalable reliability assessment paradigm for additively manufactured parts with microscopic defects, promising to advance fatigue life prediction technology for metal 3D printing in critical fields such as aerospace toward higher precision and intelligence.