Objective: To explore sex differences in 3D T1texture features in the progression of Alzheimer's disease (AD) and to predict the diagnosis of AD patients of different sex． Methods: Seventy-seven AD patients (34 males and 42 females) ，74 amnestic mild cognitive impairment ( aMCI) patients ( 35 males and 39 females) and 75 healthy controls (HC) (35 males and 40 females) were recruited and high-resolution 3-dimensional T1 structural images were collected． Brain regions closely related to AD brain damage were selected as regions of interest ( ROIs) ，texture feature extraction and feature screening were performed．Analyses were performed by sex，and the support vector machine (SVM) was used for classification and prediction. Results : In the AD vs HC，AD vs aMCI and aMCI VS HC groups by different sex，we obtained some brain regions with relatively high recognition index in different subgroups，and found that there were significant differences between female patients and male patients with high recognition index，and the recognition index of female patients ( area under the curve，accuracy，sensitivity and specificity were generally higher than that of male． Conclusion There are significant sex differences in texture features in AD process，and the classification and prediction ability of texture features in female patients is better， suggesting the importance of sex differences in AD research．This study provides some reliable biomarkers for early sex-specific identification of AD，which may be helpful for the early diagnosis and treatment of AD in the future.