Objective: Based on machine learning methods, prediction models were established to explore the predictive value of perioperative blood transfusion and ICU for hip replacement surgery in elderly patients with femoral neck fracture. Methods: The clinical data of 500 patients with femoral neck fractures undergoing hip replacement surgery at Nanjing First Hospital from January 2012 to December 2021 was analyzed.Predictive models for blood transfusion after hip replacement surgery and whether to enter the ICU were established, the predictive efficacy of different models were evaluated. The risk factors that affect the treatment in perioperative period of hip replacement surgery were explored.The receiver operating characteristic (ROC) curve was plotted,and the area under theROC curve(AUC), accuracy, sensitivity, specificity and F1 score were used to evaluate the predictive performance of the models, the importance score of the models predictive variables with the best predictive performance was obtained. Results: Using blood transfusion as the outcome variable, the AUC value, accuracy, and specificity of the random forest before balancing data were the highest among the four models; After balancing the data, the support vector machine had the highest AUC value, accuracy, specificity, and F1 score. Taking whether to enter the ICU as the outcome variable, the random forest algorithm had the highest AUC value before balancing data, and the random forest algorithm performed better; After balancing the data, the support vector machine algorithm had the highest AUC value and F1 score, performing the best. The importance score of predictive variables, based on whether or not blood transfusion was used as the outcome variable, showed that preoperative hemoglobin and creatinine were of high importance. Taking whether to enter the ICU as the outcome variable, predicting the importance score of variables, before balancing the data, preoperative hemoglobin, age, and preoperative creatinine had high importance; After balancing the data, preoperative creatinine and preoperative albumin had high importance. Conclusion: By analyzing clinical data through machine learning, the perioperative focus is on the patient’s age, preoperative hemoglobin, preoperative creatinine, and preoperative albumin. Strengthening perioperative management of hip replacement surgery can help elderly patients with femoral neck fractures recover and reduce complications.