Objective To investigate the independent influences on the occurrence of ICU Acquired Weakness(ICU-AW)in elderly patients undergoing third-and fourth-level surgeries,to construct a prediction model and draw a column-line diagram,and to internally validate the model.Methods A convenience sampling method was used to select 186 elderly patients undergoing tertiary and quaternary surgeries who were hospitalized in 3 tertiary A hospitals in Beijing from May to December 2022 as the study subjects.Single-factor and multifactor logistic regression were used to analyze the risk factors for ICU-AW in elderly patients undergoing third-and fourth-degree surgeries.A risk prediction model was established and the model was visualized by drawing a column-line diagram,and the receiver operator characteristic curve(ROC)and the Hosmer-Lemeshow tests were applied to verify the predictive effect of the model.Results ICU-AW occurred in 40 of 186 cases in the modeling group,with an incidence rate of 21.5％.The results of univariate analysis showed that the 2 groups of preoperative physiology score and surgical severity score included in the physiology and surgical severity scoring system,age,presence of cardiac disease,hemoglobin(within 24 h of admission to the ICU),blood urea nitrogen(within 24 h of admission to the ICU),blood creatinine(within 24 h of admission to the ICU),presence of braking,mode of establishment of mechanical ventilation,presence of nutritional therapy,number of sedative or analgesic drugs used,whether vasoactive drugs were used,whether diuretics were used,and the level of hemoglobin,blood urea nitrogen and blood creatinine within 24 h after admission to ICU the difference is statistically significant(P＜0.05).The results of multifactorial logistic regression analysis showed that preoperative physiology scores included in the physiology and surgical severity scoring system,the presence of cardiac disease,the presence of braking,the presence of nutritional therapy,and the number of sedative or analgesic medications used were the predictors of the occurrence of ICU-AW in elderly patients undergoing third-and fourth-degree surgeries(OR were 1.364,2.344,5.568,5.823,1.109,all P＜0.05).The above 5 factors were plotted as independent variables in a column-line graph,and the area under the ROC curve of the model was 0.859(95％CI 0.793～0.924),with an optimal critical value of 0.156,a sensitivity of 0.875,a specificity of 0.705,and a Hosmer-Lemeshow goodness-of-fit test of x2=3.906,P=0.865,Brier score of 0.109,and a decision analysis curve indicating that patients could benefit.Conclusion The predictive effect of the constructed model is good,and it can be used as a reference for early and rapid identification of the risk of ICU-AW in elderly patients undergoing third-and fourth-degree surgeries by clinical staff,and timely provision of preventive intervention programs.