Knee osteoarthritis (KOA) is a common bone disease in older patients. Medication adherence is of great significance in the prognosis of this disease. Therefore, this study analyzed the high-risk factors that lead to medication nonadherence in patients with KOA and constructed a nomogram risk prediction model. The basic information and clinical characteristics of inpatients diagnosed with KOA at the Department of Orthopedics, The Affiliated Hospital of Chengde Medical University, were collected from January 2020 to January 2022. The Chinese version of the eight-item Morisky scale was used to evaluate medication adherence. The Kellgren-Lawrence (KL) classification was performed in combination with the imaging data of patients. Least absolute shrinkage and selection operator regression analysis and logistic multivariate regression analysis were used to analyze high-risk factors leading to medication nonadherence, and a prediction model of the nomogram was constructed. The model was internally verified using bootstrap self-sampling. The index of concordance (C-index), area under the operating characteristic curve (AUC), decision curve, correction curve, and clinical impact curve were used to evaluate the model. A total of 236 patients with KOA were included in this study, and the non-adherence rate to medication was 55.08%. Seven influencing factors were included in the nomogram prediction: age, underlying diseases, diabetes, age-adjusted Charlson comorbidity index (aCCI), payment method, painkillers, and use of traditional Chinese medicine. The C-index and AUC was 0.935. The threshold probability of the decision curve analysis was 0.02-0.98. The nomogram model can be effectively applied to predict the risk of medication adherence in patients with KOA, which is helpful for medical workers to identify and predict the risk of individualized medication adherence in patients with KOA at an early stage of treatment, and then carry out early intervention.Copyright © 2023 the Author(s). Published by Wolters Kluwer Health, Inc.