Abstract: PET/MRI scanning and deep learning are of great significance in the diagnosis of Alzheimer disease (AD). The authors review the specific research progress of deep learning in AD-related PET/MRI imaging analysis and clinical diagnosis, analyze the current limitations and challenges, and point out that the combined application of deep learning and PET/MRI multimodal imaging technology significantly optimizes the imaging data processing and analysis efficiency of AD, and is a key auxiliary tool for achieving accurate diagnosis and treatment of AD in the future.