The goal of this study was to evaluate the clinical characteristics and effects of brain temperature management in patients with severe traumatic brain injury (TBI). A total of 1091 patients were registered from the Japan Neurotrauma Data Bank Project 2009. Those with a Glasgow Coma Scale (GCS) score of 9 or more, a GCS score of 3, bilateral dilated pupils, or cardiopulmonary arrest on arrival were excluded. This left a total of 401 patients. Patients were classified into three groups: no temperature management, with no intervention for brain temperature (225 patients, 56.1%), intensive normothermia (129 patients, 32.2%), and hypothermia (47 patients, 11.7%). Patient age, GCS score, pupillary abnormality, Injury Severity Score (ISS), intracranial pressure (ICP) monitoring, and outcome according to CT classification (Traumatic Coma Data Bank classification) on admission were examined. Patients were significantly older in the no temperature management group (average age 61.5 years) compared with normothermia (53.6 years) and hypothermia (46.9 years). ICP monitoring was significantly decreased in 85.1% of patients with hypothermia, 42.6% with normothermia, and 14.7% in no temperature management group. Favorable outcome rate was significantly higher with hypothermia (52.4%) compared with normothermia (26.9%) and no temperature management (20.7%) with evacuated mass lesions in contrast to diffuse injury. Multivariate analysis in patients with evacuated mass lesions showed that GCS (≥6 pts), and hypothermia were independent factors related to a favorable outcome. Appropriate thermoregulation of the brain for individual patients with various types of TBI are important.