Objective To systematically evaluate the risk prediction models of cognitive frailty in older adults in China, so as to provide a reference for nursing staff choosing or developing suitable risk prediction models of cognitive frailty in older adults, and to provide evidence for forming intervention programs. Methods The Cochrane Library, Embase, PubMed, Web of Science, CNKI, Wanfang and VIP databases were searched for relevant studies on the Chinese population from the date of inception to 1 July 2023. Literature screening was performed based on inclusion and exclusion criteria. Two researchers independently extracted data with a data extraction table and assessed the quality of studies with a bias risk assessment tool. Results A total of 10 articles were included, and 10 risk prediction models of cognitive frailty in older adults were identified, with a total sample size of 268-1 182, and 54-293 of them suffered from cognitive frailty events. One model wasn′t validated, two models were only internally validated, three models were externally validated, and four models were internally and externally validated, the area under the curve (AUC) of the model was 0.710-0.991, and eight studies calibrated the model. Independent predictors of the multivariate model included age, exercise habits, depression, comorbidities, and nutritional status. The bias risk of the included models was high, and the applicability evaluation was reasonable. Conclusion The risk prediction models of cognitive frailty for older adults in China have good discrimination and applicability, but there are significant methodological flaws and high bias risk. Future researchers should strictly follow the reporting standards for developing and evaluating the risk prediction models of cognitive frailty in Chinese older adults, and verify their feasibility in clinical practice.