Objective To evaluate the value of MRI in diagnosis of obstetrical brachial plexus palsy pre-gangl ionic injury. Methods From November 2006 to February 2008, 10 patients with obstetrical brachial plexus palsy were treated, 8 males and 2 females, aged from 2 months to 3 years (11.4 months on average). There were 7 cases of left side and 3 of right side.According to Tassin classification, 2 cases were type II, 6 type III and 2 type IV. All patients were performed MRI examinations before the operation, whose results were compared with those of exploration during the operation. Results MRI examinations showed 1 patient was normal and 9 patients had post-traumatic spinalmeningolcele. The 6 patients had displacement of spinal cord (4 towards the healthy side and 2 towards the sick side), 6 had deformity of spinal cord, and 2 had avulsed nerve root thickening. MRI detected 19 nerve roots were positive, 16 were true positive and 3 false positive in surgical exploration. MRI detected 6 nerve roots were negative, 4 were true negative and 2 false negative in surgical exploration. The sensitivity, specificity and accuracy of MRI in diagnosis of obstetrical brachial plexus palsy pre-gangl ionic injury were 84.2%, 80.0% and 83.3%, respectively. There were significant differences in the results by preoperative MRI examinations and by exploration during the operation (P lt; 0.05). Conclusion MRI can show pre-gangl ionic injury of brachial plexus of the patients with obstetrical brachial plexus palsy and can supply references for early diagnosis and operation time. MRI can be routinely conducted as a preoperative examination.