Diagnosis of developmental disorders and the assessment of brain development in pediatric patients often rely on subjective evaluations, which complicate the quantification of diagnoses. This is the first study to assess the impact of brain injuries, such as hematomas and fractures, on the accuracy of brain age estimation using existing methods, specifically a 3D-convolutional neural network (CNN) model proposed by Morita et al. The study includes 221 pediatric patients aged 0 to 3 years, with 204 healthy subjects and 17 with brain injuries. The model’s performance on pediatric brain CT images reveals significant discrepancies in brain age estimation between healthy and injured groups, with a statistically significant difference observed between the normal and hematoma groups (p-value = 0.0045). These findings underscore the importance of developing more robust brain age estimation methods that can account for brain injuries in pediatric patients. Given the clinical importance of accurate brain age estimates for diagnosing and monitoring brain development, further research is necessary to refine the model’s applicability in trauma cases and ensure its reliable use in clinical practice for pediatric trauma patients.