Background: This study aimed to inspect the association between the response to etanercept among patients with rheumatoid arthritis and several epidemiological and clinical variables and medication adherence, as measured by medication possession ratio (MPR). Methods: A cohort study that enrolls 120 active rheumatoid arthritis patients. The baseline values of disease activity score for 28 joints (DAS28) score,erythrocyte sedimmintation rate (ESR), WBC, tender joints count (TJC), swelling joints counts (SJC), and medication adherence, as measured by medication possession ratio (MPR), were identified. All patients received etanercept treatment for three months, and then the clinical response to etanercept was assessed after the end of the three months duration. Factors affecting clinical response were evaluated by univariate and multivariate logistic regression analysis. The predictive performance of a single independent predictor was then assessed using a receiver operating characteristic (ROC) curve. Results: The results of the univariate logistic regression model showed that the smoking, disease duration, baseline DAS28, and MPR could predict the patients' proclivity for being non-responder. The multivariate logistic regression model showed that only baseline DAS28 (P< 0.0001, OR=32.239, 95%CI: 4.941–210.338) and MPR (P=0.002, OR=0.00063, 95%CI: 0.00001–0.032) were independent predictive factors for the tendency of patients to be non-responder. ROC curve analysis disclosed that baseline ESR and DAS28 have a good area under the curve (AUC) with the optimal cut-off for the baseline ESR threshold was 52 mm/hr., whereas the baseline DAS28 threshold was 5.79. Conclusion: Current smoking is the main epidemiological factor that can predict the tendency for being non-responder. The potential of baseline ESR and DAS28 values as biomarkers for clinical response to etanercept in RA patients was identified by Receiver operating characteristic(ROC) analysis.