Alzheimer disease （AD） is a common neurodegenerative disease characterized by progressive cognitive impairment. Its high incidence and poor prognosis place a heavy burden on society and families. Therefore， early identification and early intervention of subjects who are at risk but have not shown symptoms is of critical significance， and magnetic resonance imaging （MRI） will play a vital role in this regard. In recent years， some machine learning （ML） methods have been proposed to apply different types of MRI features for AD conversion prediction. Multimodal MRI plays an increasingly important role in the diagnosis of AD. The research progress of AD clinical multimodal imaging is reviewed.