Summary Background: Biological age reflects physiological ageing and frequently diverges from chronological age. Patients with rheumatoid arthritis (RA) show accelerated biological ageing; however, existing indicators were developed in general populations and may not capture RA-specific pathology; no RA-specific indicator exists. We aimed to develop a biological age model and evaluate its association with ageing-related outcomes and anti rheumatic treatment.MethodsIn this retrospective observational study using two anonymised hospital-based electronic medical record databases, we included patients with RA aged ≥40 years. A biological age indicator (R-BioAge) was constructed using XGBoost with a Cox proportional hazards objective, incorporating routinely collected features. Models were trained on 60% of patients, validated on 40%, and then compared with Pheno Age and chronological age for all-cause mortality and ageing-related outcomes. Longitudinal changes after anti rheumatic drug initiation were assessed using linear mixed-effects models and group-based trajectory modelling. Findings: A total of 69,914 patients were included in the study. R-BioAge discriminated all-cause mortality better (Harrell's c-index 0·922, 95% CI 0·918–0·925) than the original PhenoAge (0·805, 0·800–0·811), refitted Pheno Age (0·816, 0·810–0·821), and chronological age (0·707, 0·699–0·714), and outperformed them for ageing-related outcomes. Key contributors included serum albumin, red cell distribution width, cumulative glucocorticoid dose, and inflammatory markers. Interleukin-6 inhibitors and abatacept were associated with reduced R-BioAge acceleration. A worsening trajectory of R-BioAge was linked with worse ageing-related outcomes.InterpretationR-BioAge, an RA-specific biological age indicator, showed strong predictive performance for multiple ageing-related outcomes, and its longitudinal trajectory was associated with anti rheumatic treatment and various clinical outcomes. R-BioAge constitutes a potential biomarker of pathological ageing to support risk stratification and treatment decisions in RA.