A clinical coma scale modified from the Glasgow Coma Scale used for humans has been suggested as a useful predictor of outcome in the head trauma patient. The objective of this study was to correlate the modified Glasgow Coma Scale (MGCS) score of dogs with head trauma with their probability of survival. Thirty-eight dogs with head trauma were selected and retrospectively evaluated. The information retrieved from the medical record of each dog included signalment, body weight, cause of head trauma, MGCS, presence of concurrent neck pain, and outcome (dead or alive) after 48 hours. Logistic regression was used to model survival in the 1st 48 hours as a function of MGCS, gender, weight, and calvarial fractures. The MGCS ranged from 5 to 18. Seven dogs died within 48 hours of the head trauma. The MGCS could predict the probability of survival in the 1st 48 hrs after head trauma with 50% probability in a patient with a score of 8. Gender, weight, and presence of skull fractures did not predict survival. In conclusion, the MGCS is a useful index for prediction of outcome in dogs with head trauma.