Objective: To develop a clinical prediction model for nonadherence to biologic treatment in rheumatoid arthritis patients. Methods: We developed a prediction model based on the training data of 102 rheumatoid arthritis patients, and the data were collected from January 2018 to March 2019. Adherence was evaluated using the proportion of treatment days covered within 6 months as the outcome event. The LASSO regression model was used to optimize feature selection for medication nonadherence risk model. Multivariate Logistic regression analysis was applied to build a predicting model incorporating the feature selected in LASSO regression model. Predictability, calibration, and clinical usefulness of the predicting model were assessed using the C-index, calibration plot, ROC curve, decision curve analysis and internal validation was assessed using Bootstrap. Results: Predictors in the prediction model included age, disease activity, education level, monthly income and anxiety degree. The model displayed good prediction ability; C index was 0.897 (95%CI: 0.827–0.906), and the area under the ROC curve was 0.8787939 with a good calibration ability. High C index of 0.888 was reached in internal verification. Decision curve analysis showed that the prediction effect of the model could benefit about 85% of the patients without harming the interests of other patients. Conclusion: The clinical prediction model could help clinical medical staff to identify patients with higher risk of nonadherence early, so that intervention measures can be taken in time.