Objective To explore the effect of controlled release of nerve growth factor (NGF) on peripheral nerve defect repaire by acellular nerve graft. Methods The microspheres of NGF were prepared with drug microsphere technologyand fixed with the fibrin glue to make the compl icated controlled release NGF. Twenty healthy male SD rats weighing 280-300 g were adopted to prepare acellular xenogenous nerve, 52 male Wistar rats weighing 250-300 g were adopted to prepare the 10 mm defect model of left sciatic nerve. and thereafter were randomly divided into 4 groups: autograft group(group A), acellular nerve allograft combined with the double controlled release NGF (group B), acellular nerve allograft (group C) and acellular nerve allograft combined with fibrin glue (group D). Without any operation, the right sciatic nerve was regarded as control group. General observation was conducted after operation. The nerve axon regeneration length was measured 2 weeks after operation. The effects of peripheral nerve regeneration were evaluated by neural electrophysiology, the recovery rate of triceps surae muscular tension and weight and histological assessment 16 weeks after operation. Results All the animals survived till the end of experiment. The length of nerve regeneration was measured at 2 weeks after transplantation. The regeneration nerve of group A was longer than that of other groups (P lt; 0.05), group B longer than groups C and D (P lt; 0.05), and there were no difference between group C and group D (P gt; 0.05). At 16 weeks after operation, the recovery rates of nerve conduction velocity of groups A and B (73.37% ± 7.82% and 70.39% ± 8.45%) were larger than that of groups C and D (53.51% ± 6.31% and 55.28% ± 5.37%) (P lt; 0.05). The recovery rates of the triceps surae muscular tension in group A (85.33% ± 5.59%) were larger than that in groups B, C and D (69.79% ± 5.31%, 64.46% ± 8.49％ and 63.35% ± 6.40%) (P lt; 0.05). There were no significant differences among groups B, C and D (P gt; 0.05). The recovery rates of the triceps surae weight in group A (62.54% ± 8.25%) werelarger than that in groups B, C and D (53.73% ± 4.56%, 46.37% ± 5.68% and 45.78% ± 7.14%, P lt; 0.05). There was significant difference between group B and groups C, D (P lt; 0.05) and no significant differences between group C and group D (P gt; 0.05). The histological observation indicated that axon number and myel in thickness in group B were larger than those in group C and group D (P lt; 0.05). The axonal diameter in group B was significantly less than that in group A (P lt; 0.05). Conclusion Acellular nerve graft combined with the controlled release NGF is a satisfactory alternative to repair the peripheral nerve defect.