Objective: To analyze the indications and treatment outcomes of decompressive craniectomy (DC) for patients with severe traumatic brain injury (sTBI) complicated with cerebral herniation, and evaluate potential prognostic factors. Methods: Retrospective investigation of sTBI patients who underwent craniotomy surgery in our hospital from May 2022 to December 2024, including post resuscitation Glasgow Coma Scale (GCS) scores, Trauma Severity Scale (ISS) scores, pupil status, demographic and sociological data, injury mechanisms, computed tomography (CT) scans, surgical procedures, intraoperative findings, treatment outcomes, and injury to hospital time. All patients' CT scans upon admission were analyzed based on Marshall CT grading and Fisher scale, and evaluated for midline shift (MLS) and basal ganglia status. If the brain parenchyma above the inner plate of the skull protrudes during surgery, indicating brain swelling and elevated ICP, and abnormal cerebral blood flow, the doctor decides to perform controlled DC. The Glasgow Outcome Scale (GOS-E) is used to evaluate the prognosis at 6 months. Results: A total of 90 cases (59.21%) of patients received controlled DC (DC group), and the other 62 cases (40.79%) only received craniotomy (non-DC group). Compared with the non-DC group, the incidence of cerebral infarction in the DC group was significantly higher (P?＜0.001), the proportion of patients receiving ultra-early surgery was lower (P=0.007), and the GCS score at admission was lower (P=0.015). The midline displacement (MLS) was more obvious (P?＜0.001), and the median intracranial pressure (ICP) was higher (P=0.026), as well as the proportion of patients with subdural hematoma (SDH), the proportion of patients with Marshall CT grade V, and the proportion of patients with bilateral pupil dilation at admission were higher (P?＜0.05). The time from injury to admission (OR=4.688, P=0.021), hematoma type (P?＜0.05), and MLS (OR=1.177, P=0.002) were independent preoperative factors significantly associated with DC. The mortality rate within 6 months after surgery in the DC group was higher than that in the non-DC group (37.78% vs. 22.58%, P=0.048), and the median GOS-E score at 6 months after surgery was also slightly lower than that in the non-DC group (4.0 points vs. 5.0, P=0.029). There was no significant difference in the risk of major perioperative complications between the two groups (P＞0.05). After multivariate Logistic regression analysis, a lower admission GCS score, bilateral pupil dilation, time from injury to admission≥1 hour, preoperative ICP＞40 mmHg, and advanced age were independent risk factors for poor prognosis (P?＜0.05). Conclusion: Preoperative analysis of the time from injury to admission, hematoma type, and MLS in patients with sTBI accompanied by brain herniation can have guiding value for whether DC is performed subsequently. In addition, a lower admission GCS score, bilateral pupil dilation, time from injury to admission ≥1 hour, preoperative ICP＞40 mmHg, and advanced age are associated with a poor prognosis for patients.