Abstract: Objective To compare the disease activity and therapeutic strategy of late-onset rheumatoid arthritis （LORA）with young-onset rheumatoid arthritis （YORA）. Methods Medical records of 259 patients with rheumatoid arthritis (RA) were reviewed retrospectively. The cut-off between LORA and YORA was operationally set at 60 years of age at disease onset. Gender ratio, disease duration, feature of involved joint, extra-articular features, laboratory features, disease activity score and treatment strategy were compared between LORA and YORA. Results The LORA patients had a gender ratio (male/female) of 1/1.88, which was approaching 1/1 in the older age group. It took longer for LORA to be diagnosed than YORA (\P<0.001). LORA had more frequent involvement of shoulders (\P<0.001), while elbow, wrist, metacarpophalangeal joint （MCP），proximal interphalangeal joint （PIP） and ankle joints were more common in YORA (\P<0.001). LORA patients were more likely to have anemia than YORA patients (\P<0.05). No significant differences were found between LORA and YORA in specific serologic index, including rheumatoid factor and anti-cyclic citrullinated peptide antibody, and disease activity score 28-C-reactive protein （DAS28-CRP） and clinical disease activity index （CDAI）. But LORA had a higher simplified disease activity index （SDAI） (\P=0.002). Glucorcoticoid was used in 67.4% LORA patients, compared with 29.3% in YORA patients (\P<0.001). In contrast, disease-modifying anti-rheumatic drugs（DMARDs） were used in 73.7% LORA patients, compared with 97.6% in YORA patients (\P<0.001). Conclusion Delayed diagnose of LORA is common due to atypical presentation at disease onset. RA should be considered in elderly patients with large joints for differential diagnosis. LORA is more likely to have anemia than YORA, albeit no significant differences in serological index and extra-articular presentations. LORA patients should be treated with DMARDs as aggressively as YORA patients, if their comorbidities allow to do so.