ObjectiveTo investigate the risk factors associated with cognitive frailty among older adults in nursing homes using logistic regression and decision tree modeling, and to compare the predictive performance of these methods.MethodsA cross-sectional study was conducted involving 697 participants aged 60 and older residing in eight nursing homes in Sichuan province, China. Participants were recruited using convenience sampling. Data were collected through questionnaires administered to the older adults. Logistic regression and decision tree modeling were employed to construct models predicting cognitive frailty.ResultsLogistic regression analysis identified age, education degree, exercise, intellectual activities, number of chronic diseases, nutritional status, sleep quality, and depression as significant predictors of cognitive frailty (all p < 0.05). The final decision tree model consisted of three layers and 17 nodes. Six factors were identified as significant predictors: sleep quality, number of chronic diseases, depression, education level, nutrition, and exercise. Receiver operating characteristic (ROC) curve analysis revealed that the area under the curve (AUC) for the logistic regression model was 0.735 (95% CI: 0.701–0.767) with a sensitivity of 0.58 and specificity of 0.75. The AUC for the decision tree model was 0.746 (95% CI: 0.712–0.778) with a sensitivity of 0.68 and specificity of 0.70.ConclusionAge, education level, exercise, intellectual activities, sleep quality, number of chronic diseases, nutritional status, and depression are significant risk factors for cognitive frailty in older adults residing in nursing homes. Both logistic regression and decision tree models demonstrated comparable predictive performance, with each offering distinct advantages. The combined use of these methods can enhance predictive accuracy and provide valuable insights for clinical practice and policy development.