Objective To determine predictors and optimal duration of sustained remission (SR) in patients with rheumatoid arthritis (RA). Methods A total of 428 consecutive patients with RA visiting our clinic routinely between 2012 and 2013 were evaluated. Seventy seven of these patients in DAS28 remission were enrolled and followed up for 62.2 +/- 9.9 months. Patients in remission >= 6 months (SR) and shorter (non: N-SR) were compared in terms of demographic-clinical data and the psychosocial factors. At enrollment, 1st and 5th years, patients in DAS28, SDAI, and Boolean remission were determined. Results Sixty three patients were in SR and 14 in N-SR. Lower baseline DAS28 and HAQ scores, anti-CCP were positive predictors of SR. Although the presence of anxiety, depression, fibromyalgia, and fatigue were lower in the SR group, there was no significance. Patients in DAS28 remission (100%) at baseline reduced to 64% at 1st and 42.6% at 5th years. Patients satisfying SDAI and Boolean remission at these three visits were 49%, 44%, and 32.4% vs 41%, 28%, and 20.6%, respectively. If the duration of remission is defined as 6 months, the remission rates of SDAI at inclusion and fifth years' visits were similar but Boolean remission rates differed significantly and if it is accepted as >= 12 months, both the SDAI and Boolean remission rates were not different. Conclusion Low DAS28 and HAQ scores at baseline, anti-CCP were positive predictors of SR. Instead of 6 months, remission duration for >= 12 months would probably help us to predict SR independently from the chosen criteria; Boolean or SDAI.