The study aimed to assess in a population of subjects with rheumatoid arthritis (RA) treated with methotrexate (MTX) how the initial approach to the treatment influenced subsequent disability. We performed a cross-sectional analysis of data collected during the baseline visit of the MARI study, a multicenter observational study on patients with RA on treatment with MTX for at least 12 months. Subjects who fulfilled the Health Assessment Questionnaire (HAQ) were included in the evaluation. For every patient we retrospectively evaluated the disease duration, the duration of symptoms before the diagnosis, the time elapsed before first MTX treatment, the initial MTX dose, and the concomitant medications in the first six months of therapy. Disability was defined as a DI-HAQ score ≥1. The study population included 1015 subjects. Patients with a DI-HAQ score ≥1 had a longer duration of symptoms before diagnosis, a higher delay in treatment initiation, a lower initial dose of MTX and a more frequent co-treatment with symptomatic drugs. Disability was found less frequently in subjects treated with other concomitant disease modifying anti-rheumatic drugs (DMARDs) but not with biological agents. Logistic regression analysis identified as significant predictors of disability: older age, female sex, a longer time to complete diagnosis, a delay in starting MTX treatment higher than 6 months, and a concomitant treatment with symptomatic drugs, while a combination therapy with other DMARDs was associated with a lower risk of disability. A late diagnosis and a delay in starting a treatment with MTX are associated with poorer functional outcomes in patients with RA.