Objective To construct adenovirus expressing NGF (Ad-NGF) and to investigate its promotive effect on the reparation and regeneration of sciatic nerve injury in rats. Methods NGF gene sequence was cloned into shuttle plasmid pCA13 of adenovirus type 5. After packed in HEK-293 cells, the recombinant adenoviruses-Ad-NGF underwent sequence identification. Thirty-two male SD rats weighing 180-200 g were randomly divided into 4 groups (n=8 rats per group). Sciatic nerve injury model was establ ished by disconnecting and direct suturing the right sciatic nerve in the rat. Theright gastrocnemius muscle of group A and C received Ad-NGF injection and adenovirus vector without NGF gene sequence injection, respectively, and 1 × 108 PFU/per time was given every other day for three times. Group B and D received NGF injection (200 U/d) and normal sal ine (100 ?L/d), respectively, for 3 weeks. The effect of various treatments on injured sciatic nerve was evaluated by performing sciatic nerve function index and nerve electrophysiology detections 31 days after operation. Meanwhile, the sciatic nerve in the anastomosis and at the site 1 cm distal to the anastomosis were obtained, and underwent RTPCR and Western blot analysis for detecting NGF mRNA and protein expression level in the injured sciatic nerve in the rats. Histology, immunohistochemistry, and transmission electron microscope observations were conducted. Results Ad-NGF carrying NGF gene sequence was constructed successfully and confirmed by sequence analysis. The sciatic nerve function index, nerve conduction velocity, evoked potential ampl itude, and latent period of group A was better than those of other groups (P lt; 0.05), and there were no significant differences among group B, C, and D (P gt; 0.05). RT-PCR and Western blot detection: the expression levels of NGF mRNA and protein in group A were greater than those of group B, C, and D (P lt; 0.05), and no significant differences were noted among group B, C, and D (P gt; 0.05). Histology and immunohistochemistry observation showed that the regeneration of the sciatic nerve in group A was obvious superior to that of other groups. Transmission electron microscopy observation suggested there was significant difference between group A and groups B, C, and D in terms of axonal diameter of sciatic nerve cross-section, myel in sheath thickness and nerve fiber number (P lt; 0.05), and there were no significant differences among group B, C, and D (P gt; 0.05). Conclusion Ad-NGF can effectively promote the repair of sciatic nerveinjury in rats, and is a new method for obtaining large amounts of NGF in the area of injured peripheral nerve.