Objective: This study aims to clarify the relevant factors influencing practitioners' methods of prescribing medications for bipolar disorder, in a nationwide survey in Japan. Methods: The clinical records of 3130 outpatients with bipolar disorder were consecutively reviewed from 176 psychiatric outpatient clinics. Fifteen parameters, that is, five patients' including five general characteristics (sex, age, education, occupation, and social adjustment), five patients' aspects of mental functioning (onset age, comorbid mental illness, rapid-cycling, psychopathologic severity, and followed-up years), and five practitioners' characteristics (sex, age, specialist experience, clinic standing years, and location), were evaluated. The number of psychotropic drugs (mood stabilizers, antidepressants, antipsychotic drugs, anxiolytics, and hypnotics) was used as an index of pharmacotherapy. Converted data from each practitioner-unit were analyzed. Results: Seven factors (patient's social adjustment, patient's psychopathology, patient's comorbid mental disorders, patient's followed-up years, doctor's age, clinic running years, and patient's education years) were correlated to the number of psychotropic drugs. Multiple regression analysis showed that the severity of illness (poor social adjustment, and comorbid mental illness) and an intractable disease course (long followed-up years), were significantly associated with the number of psychotropic drugs. Conclusion: Our findings indicated that patient-related conditions affected psychotropic polypharmacy more strongly than did practitioner-related conditions.