Rheumatoid arthritis (RA) is a chronic inflammatory disease that is very complex and heterogeneous. If not adequately treated, RA patients are likely to manifest excess of morbidity and disability with an important impact on the quality of life. Pharmacological treatment is based on the administration of the disease-modifying antirheumatic drugs (DMARDs), subdivided into conventional synthetic (csDMARDs), targeted synthetic (tsDMARDs), and biological (bDMARDs). bDMARDs are now frequently administered in patients, both as alternative treatment and together with csDMARDs. Unfortunately, there is a therapeutic response variability both to old and new drugs. Therefore, to identify pre-therapeutic and on-treatment predictors of response is a priority. This review aims to summarize recent advances in understanding the causes of the variability in treatment response in RA, with particular attention to predictive potential of autoantibodies and DMARD pharmacogenetics. In recent years, several biomarkers have been proposed to personalize the therapy. Unfortunately, a magic bullet does not exist, as many factors concur to disease susceptibility and treatment outcomes, acting around the patient’s congenital background. Models integrating demographic, clinical, biochemical, and genetic data are needed to enhance the predictive capacity of specific factors singularly considered to optimize RA treatment in light of multidisciplinary patient management.