Objective To analyze the prevalence and influencing factors of subsyndromal delirium (SSD) in elderly patients after elective hip or knee replacement, to construct a risk prediction model, and to provide reference for prevention and management of SSD. Methods This was a prospective cohort study of 415 patients undergoing elective hip or knee replacement in a single tertiary hospital. LASSO regression and multivariate logistic regression were used to analyze the influencing factors. The nomogram prediction model was constructed using R software and validation was performed in 121 patients. Results The incidence of SSD in model development group was 42.41%. The Mini-Mental State Examination score before the surgery (OR=0.870), transfer to ICU after the surgery (OR=3.965), postoperative C-reactive protein (OR=1.020), and postoperative pain score (OR=2.642) were factors associated with SSD (all P<0.05). The nomogram predication model had good performance, with the area under the ROC curve (AUC) of 0.831, sensitivity of 0.716, and specificity of 0.824. The AUC of Bootstrap internal validation and of validation group was 0.826 and 0.812 respectively. The clinical decision curve showed that the model had good clinical net benefit. Conclusion SSD is prevalent in elderly patients undergoing elective hip or knee replacement, and the nomogram prediction model has satisfactory predictive efficacy, which facilitates medical staff to develop targeted intervention.