Objectives: The main objective is to study the contribution of illness and medication beliefs to treatment adherence in patients with rheumatoid arthritis. Methods: the design was a cross-sectional study. The compliance Questionnaire for Rheumatology (CQR) was used to measure therapeutic adherence. The beliefs about medicines questionnaire (BMQ) and the brief illness perception questionnaire (IPQ-b) were used to assess patient’s beliefs about medicines and about the disease. Other factors studied were treatment satisfaction, patient´s demographic and clinical characteristics. Results: 144 patients were included in the study, 113 (78.4%) patients showed good treatment adherence. Patients with poor adherence presented higher scores in the BMQ harm domain (13±5 vs. 11±3, p= 0.013). Meanwhile, patients with good adherence presented higher scores in the necessity BMQ domain (21±3 vs. 20±3, p= 0.015), increased feeling of treatment control (8.8± 1.5 vs. 7.7± 2.1,p= 0.008), higher emotional response (6.2±3.1 vs. 4.8±3.4,p= 0.042) and a higher level of treatment satisfaction (77.2±12.4 vs. 69.9±12.5,p=0.004). In a multivariate analysis for each unit of increase in the score of BMQ´s harm domain, adherence was reduced by 20% (CI 95% 0.08-0.3, p= 0.001); for each unit of increase in treatment control item of the IPQ-b, adherence increased 1.4 times (CI 95% 1.1-1.8,p= 0.006); and for each unit of increase in the emotional response item of the IPQ-b, adherence increased 1.3 times (CI 95% 1.1-1.5,p= 0.002). Conclusion: in our cohort of RA patients, good adherence is associated with stronger treatment necessity perception, stronger feeling of treatment control, higher emotional response and higher level of treatment satisfaction; on the other side, patients with poor adherence had stronger beliefs of medicines as harmful substances