Objective To investigate the therapeutic efficacy of endoscopic optic nerve decompression (EOND) on patients with traumatic optic neuropathy (TON) and their related prognostic factors. Methods The clinical data of 103 eyes with TON,admitted to our hospital from January 1995 to February 2001 (n=53) and from March 2001 to December 2009 (n=50),were retrospectively analyzed; the clinical outcomes of these patients after being performed EOND were observed. Univariate analysis (Chi-square test) and logistic regression were performed to analyze the prognostic factors of TON underwent EOND therapy. Results The total effective rate of 103 eyes was 37.86％ (39/103); the effective rate of eyes with residual vision was 83.3％ (20/24),which was significantly higher than that of those without residual vision (24.05％,19/79,P＜0.05).Univariate analysis and multiple factor logistic regression,respectively,showed that 3 variables significantly increased the risk of the visual acuity:no residual vision, interval from injury to surgery for more than 3 d and ethmoid and/or sphenoid sinus hematocele after injury.Other factors such as age,disturbance of consciousness,optic canal fracture,performing of sheathotomy,hormone level and interval from injury to surgery for more than 7 d were not correlated with the efficacy (P＞0.05).Comparing fractures at different sites and with different severities,the efficacy of surgery was also different. The logistic regression equation is P (1)=1/ [1+e-(-2139+2.839X3+1.372X5+2.263X9)]. Conclusion The total therapeutic efficacy of EOND on patients with TON is not satisfactory, especially for those without light perception. The prognostic factors of TON were very complex and interactive; no light perception, interval from injury to surgery for more than 3 d,hemorrhage within the ethmoid and/or sphenoid sinus and lateral or multiple fracture of optic canal are factors affecting the prognosis of EOND for patients with TON. The indication for EOND must be individualized.