Purpose/Significance To construct an intelligent auxiliary diagnosis model，so as to enhance the early diagnosis effi-ciency and clinical decision support level of knee osteoarthritis. Method/Process The magnetic resonance imaging(MRI)image dataand clinical information of the knee joint of patients suspected of having knee osteoarthritis in Tangshan Second Hospital are retrospec-tively collected. A standardized dataset is established，and an intelligent auxiliary diagnosis model is constructed using convolutional neural network. The model’s efficacy is systematically evaluated by using methods such as the receiver operating characteristic curve， confusion matrix and Kappa consistency test，etc. Result/Conclusion This model converts deep learning technology into an auxiliarydecision-making tool for clinical imaging. Its diagn ostic efficacy is higher than that of routine clinical diagnosis by physicians，which is conducive to enhance diagnostic efficiency and optimize the medical treatment process.