Purpose/Significance To discuss the application value of deep learning (DL) technology in the automatic segmentation of left atrial CTA images, to review the theoretical framework and research progress at home and abroad, and to sort out the potential research directions. Method/Process Domestic and foreign literatures are retrieved. The traditional segmentation methods of left atrial CTA images and the algorithms based on DL are compared and analyzed, and their effects are evaluated. Result/Conclusion DL technology has significantly improved the segmentation accuracy of left atrial images, providing strong support for clinical diagnosis and treatment. However, there are still research gaps that need to be further explored.