The aim of this study is to explore relationship between HLA-DRB1 alleles and the susceptibility and clinical features of rheumatoid arthritis (RA) in the south Tunisian population. We studied 142 RA patients and 123 controls matched for age, sex, and ethnicity. HLA-DRB1 genotyping and HLA-DRB1*04 subtypes were performed using polymerase chain reaction/sequence-specific primers. Association was assessed based on the χ (2) test and odds ratio with 95% confidence interval. For multiple comparisons, p value was corrected (p (c)) with Bonferroni test. Two alleles, HLA-DRB1*04 (p=0.045, p(c)=NS) and HLA-DRB1*10 (p=0.021, p(c)=NS), were found to have increased frequencies in RA patients compared to controls. In contrast HLA-DRB1*08 allele was found to have a decreased frequency in patients compared to controls (p=0.044, p(c)=NS). Molecular subtyping of the most prevalent allele (DRB1*04) revealed increased frequencies of HLA-DRB1*04:05 in patients compared to controls (p=0.013, p(c)=NS) whereas HLA-DRB1*04:02 showed a protective effect (p=0.005, p(c)=0.04). Moreover, stratified analyses indicated statistically significant associations between HLA-DRB1*04 allele and anti-cyclic peptides antibodies positivity (ACPA(+)) and rheumatoid factor positivity (RF(+); p(c)=0.03, for both subgroups), HLA-DRBI*10 and ACPA(+) and the presence of another autoimmune disease (p(c)=0.05 and p(c)=0.007, respectively), and HLA-DRB1*04:05 and RF(+) and erosion (p(c)=0.005 and p(c)=0.049; respectively). A significant decrease in the frequency of the DRB1*04:02 allele was observed in patients with ACPA(+) and RF(+) subgroups (p(c)=0.04 and p(c)=0.02, respectively). Our results showed that there was a trend of positive association of HLA-DRB1*04 and HLA-DRB1*10 with RA as such and significant associations with the disease severity in the south Tunisian population.