Tissue-engineered muscle has been proposed as a solution to repair volumetric muscle defects and to restore muscle function. To achieve functional recovery, engineered muscle tissue requires integration of the host nerve. In this study, we investigated whether denervated muscle, which is analogous to tissue-engineered muscle tissue, can be reinnervated and can recover muscle function using an in vivo model of denervation followed by neurotization. The outcomes of this investigation may provide insights on the ability of tissue-engineered muscle to integrate with the host nerve and acquire normal muscle function. Eighty Lewis rats were classified into three groups: a normal control group (n = 16); a denervated group in which sciatic innervations to the gastrocnemius muscle were disrupted (n = 32); and a transplantation group in which the denervated gastrocnemius was repaired with a common peroneal nerve graft into the muscle (n = 32). Neurofunctional behavior, including extensor postural thrust (EPT), withdrawal reflex latency (WRL), and compound muscle action potential (CMAP), as well as histological evaluations using alpha-bungarotoxin and anti-NF-200 were performed at 2, 4, 8, and 12 weeks (n = 8) after surgery. We found that EPT was improved by transplantation of the nerve grafts, but the EPT values in the transplanted animals at 12 weeks postsurgery were still significantly lower than those measured for the normal control group at 4 weeks (EPT, 155.0 -38.9 vs. 26.3 -13.8 g, p < 0.001; WRL, 2.7 -2.30 vs. 8.3 -5.5 s, p = 0.027). In addition, CMAP latency and amplitude significantly improved with time after surgery in the transplantation group ( p < 0.001, one-way analysis of variance), and at 12 weeks postsurgery, CMAP latency and amplitude were not statistically different from normal control values (latency, 0.9 -0.0 vs. 1.3 -0.7 ms, p = 0.164; amplitude, 30.2 -7.0 vs. 46.4 -26.9 mV, p = 0.184). Histologically, regeneration of neuromuscular junctions was seen in the transplantation group. This study indicates that transplanted nerve tissue is able to regenerate neuromuscular junctions within denervated muscle, and thus the muscle can recover partial function. However, the function of the denervated muscle remains in the subnormal range even at 12 weeks after direct nerve transplantation. These results suggest that tissue-engineered muscle, which is similarly denervated, could be innervated and become functional in vivo if it is properly integrated with the host nerve.