目的 对国内外缺血性脑卒中后认知障碍风险预测模型进行范围综述，为临床护理实践及后续研究提供参考与借鉴。 方法 全面检索9个中英文数据库，检索时限为建库至2025年3月2日，提取缺血性脑卒中后认知障碍的发生率、模型纳入的预测因子、建模方法及性能指标并进行汇总分析。 结果 共纳入24项研究，缺血性脑卒中后认知障碍的发生率为30.4%~57.4%。每项研究纳入的预测因子数目为1~17个不等，其中年龄和教育水平为出现频次最高的预测因子。模型的构建方式为Logistic回归和机器学习。纳入模型的受试者工作特征曲线下面积（Area Under the Curve，AUC）为0.65~0.98。 结论 现有模型普遍具有高偏倚风险且性能差异大，未来应通过合理纳入预测因子、妥善处理缺失数据、加强内部验证和外部验证等途径综合提升模型质量。

Objective To conduct a scoping review of risk prediction models for cognitive impairment after ischemic stroke both domestically and internationally，aiming to provide references for clinical nursing practice and subsequent research. Methods A comprehensive search was conducted in 9 Chinese and English databases from the establishment of the databases to March 2，2025. The incidence of cognitive impairment after ischemic stroke，the predictive factors included in the models，the modeling methods，and performance indicators were extracted and summarized. Results A total of 24 studies were included. The incidence of cognitive impairment after ischemic stroke ranged from 30.4% to 57.4%. The number of predictive factors included in each study ranged from 1 to 17，with age and education level were the most frequently included factors. The modeling methods were logistic regression and machine learning. The area under the curve（AUC） values of the included models ranged from 0.65 to 0.98. Conclusion The existing models generally have a high risk of bias and significant performance differences. In the future，the quality of models should be comprehensively enhanced through reasonable inclusion of predictive factors，proper handling of missing data，and strengthening of internal validation and external validation.