The aim of the present work was to study the validity and prognostic accuracy of scores for assessing the severity of the condition in children with severe trauma, located in the Department of Anesthesiology and Resuscitation in the Clinical and Research Institute of Urgent Pediatric Surgery and Trauma. The prospective study was conducted using clinical and physiological data collected at the admission and during the first 24 hours of hospitalization from 474 patients. The validity and prognostic accuracy of prognostic scores were assessed by determining their discrimination and calibration ability. A comparison of the discriminatory ability of scores was carried out by comparing the areas under the ROC curves with the z-criterion. Four prognostic scores were included into the study: PRISM, APACHE II, ISS-RTS-TRISS, which were used for calculating the severity of injury and for prognosis of death. Score PTS was used for evaluating the severity index only. Our results indicate that only score ISS-RTS-TRISS may be useful in practice (has excellent discrimination ability and significant calibration ability). The other lack either discrimination ability (PRISM) or calibration ability (PTS, APACHE II). The result of the study has shown that only one of the four prognostic scores, ISS-RTS-TRISS, can be successfully used in everyday practice in the department of anesthesiology and resuscitation in the specialized hospital of children's traumatology to assess the severity of the condition, with the possibility of predicting the likelihood of a lethal outcome.