The aim of this study was to asses the effects on pain, disability, and health status of an educational-behavioral joint protection program in a group of moderate-severe rheumatoid arthritis (RA) patients. Eighty-five subjects with RA in treatment with anti-tumor necrosis factor alpha (TNFalpha) drugs (infliximab) were enrolled into the study and randomized into either an experimental group (46, EG) or a control group (39, CG). We organized four EG meetings, which included information on pathophysiology and evolution of RA, joint protection during normal activities of daily living, suggestions on how to adapt the surrounding environment, and self-learning exercises to perform at home. Sociodemographic characteristics and degree of knowledge of the disease, measured by the Health Service Interview (HSI), were recorded at baseline. The outcome measures included the Visual Analogue Scale (VAS), the Arthritis Impact Measurement Scale 2 (AIMS2), and the Health Assessment Questionnaire (HAQ), which were administered at the beginning and end of the trial. Thirty-six patients from the EG (7 men and 29 women; mean age 54.2 years) and 34 from the CG (6 men and 28 women; mean age 52.2 years) completed the trial. No statistical differences in baseline evaluations were found between the two groups. According to the answers given on the HSI, the majority of our patients had poor knowledge of RA and its consequences. After a mean time of 8 months, the patients receiving educational training displayed a significant decrease, compared to the CG, in the VAS (p = 0.001), HAQ (p = 0.000), and physical (p =0.000), symptoms (p = 0.049), and social interaction (p = 0.045) scores on the AIMS2, but not in other items. Our study showed that 8 months after attending an educational-behavioral joint protection program, subjects with moderate-severe RA presented less pain and disability and thus an enhanced health status. This approach may efficiently complement drug therapy in these patients.