DSM-5 diagnostic categories, defined by a set of psychopathological symptoms are heterogeneous conditions that may include different biological entities, with distinct etiopathogenesis, different courses and requiring different treatment management. For bipolar disorder the major evidences for this lack of validity are the long paths before a proper diagnosis, the inconsistence of treatment guidelines, the long phases of pharmacological adjustment and the low average of long-term treatment response rates. Personalized medicine for mental disorders aims to couple established clinical-pathological indexes with new molecular profiling to create diagnostic, prognostic and therapeutic strategies precisely tailored to each patient. Regarding bipolar disorder, the clinical history and presentation are still the most reliable markers in stratifying patients and guiding therapeutic management, despite the research goes to great lengths to develop new neuropsychological or biological markers that can reliably predict individual therapy effectiveness. We provide an overview of the advancements in personalized medicine in bipolar disorder, with particular attention to how psychopathology, age at onset, comorbidity, course and staging, genetic and epigenetic, imaging and biomarkers can influence treatment management and provide an integration to the conventional treatment guidelines. This approach may offer a new and rational path for the development of treatments for targeted subgroups of patients with bipolar disorder.