Objective investigate the changes in the expression levels of ferroptosis-related glutathione （GSH）， glutathione peroxidase 4 （GPX4）， homocysteine （Hcy）， and serum iron in the serum of patients with Alzheimer's disease （AD）， analyze their correlation with cognitive function， and evaluate the diagnostic value of serum ferroptosis-related indicators for AD. Methods Thirty-seven patients with AD hospitalized in the Second Affiliated Hospital of Xinjiang Medical University and 35 individuals without cognitive impairment hospitalized during the same period were selected. Neuropsychological assessments were conducted on all patients in the AD group and the control group， and the levels of serum GSH， glutathione peroxidase 4 （GPX4）， homocysteine （Hcy）， and serum iron were detected. The general data of the two groups were compared. The correlations between GSH， GPX4， Hcy， serum iron， and the scores of the Mini-Mental State Examination （MMSE） and the Montreal Cognitive Assessment （MoCA） were analyzed. Univariate and multivariate binary Logistic regression analyses were employed to explore the influencing factors of AD and construct its predictive model. The receiver operating characteristic （ROC） curve was plotted to evaluate the diagnostic value of the prediction model for AD. Results （1） When compared to the control group， the serum levels of GSH （23.573 ± 8.862 μmol/L） and GPX4 （21.767 ± 9.609 μU/mL） in the AD group were significantly lower （P < 0.05）. In the control group， the corresponding levels were GSH （34.380 ± 19.673 μmol/L） and GPX4 （31.345 ± 10.553 μU/mL）. There was no statistically significant difference in the serum levels of iron and Hcy between the two groups （P > 0.05）. （2） Serum GPX4 was positively correlated with MMSE （r = 0.233， P = 0.049） and MoCA （r = 0.296， P = 0.011）， and serum iron was positively correlated with MoCA （r = 0.252， P = 0.033）. （3） Multivariate binary Logistic regression analysis indicated that body mass index （BMI） （OR = 0.794， P < 0.05） and serum GPX4 （OR = 0.917， P < 0.05） were protective factors for the onset of AD. The area under the curve （AUC） for diagnosing AD using serum GPX4 levels was 0.766 （95%CI： 0.656 ~ 0.877）， with a sensitivity of 67.6% and a specificity of 77.1%， and the cut - off value was 44.7 （P < 0.05）. The AUC value for diagnosing AD by combining BMI and GPX4 levels was 0.815 （95%CI： 0.715 ~ 0.914）， with a sensitivity of 81.1% and a specificity of 77.1% （P < 0.05）. Conclusions （1） The expression levels of serum GSH and GPX4 vary between AD patients and the control group. Specifically， these levels are lower in AD patients than in the control group. （2） Serum GPX4 and serum iron are positively correlated with cognitive function. （3） BMI and serum GPX4 are influencing factors of AD， and the prediction model constructed by their combined factors has a certain predictive value for AD.