OBJECTIVE: To investigate a revision of the Trauma and Injury Severity Score (TRISS) weight coefficients in order to overcome the inference from foreign coefficients on Chinese trauma scoring. METHODS: The data of 1 297 Chinese trauma patients were studied for trauma scoring with the Revised Trauma Score-Injury Severity Score-TRISS (RTS-ISS-TRISS) system to get a serial of new weight coefficients through analyzing a multivariation logical regression between the outcome and the injury severity. RESULTS: ISS was higher but the Age Score and probability of survival (Ps) of the death group were lower than those of the survival group. New values of RTS-ISS-Age coefficients differed from the Major Trauma Outcome Study (MTOS) ones, through which the constant b(0) decreased its negative value, and ISS weight b(2) increased its negative value, but RTS weight b(1) and age weight b(3) changed with the trauma types. MTOS's values and new values of weight coefficients were used on 1297 patients for prognosis by calculating Ps. The accuracy of new values (90.13%) was a little higher than that of MTOS's (89.5%), with a promotion in specialization but a loss in sensitivity. CONCLUSIONS: A revision of TRISS's weight coefficients is either necessary or feasible. To achieve this purpose, a Chinese trauma database with large capacity is recommended.