Non-vascularized free composite graft is one of the simple and effective reconstructive options, but its clinical use has been limited due to questionable survival rate. Early vascularization is essential for graft survival and is mainly carried out via recipient bed or repaired sites. This study was designed to investigate the effect of the lateral marginal approximations on the survival of the free composite flap using a model of skin-subcutaneous composite graft in rats. Thirty 1.5 ${ imes}$ 1.5 $cm^2$ sized square shape composite flaps were elevated freely and reposed in place immediately on the dorsum of five Sprague-Dawley rats, and divided into five groups of six flaps. In all groups, graft bed was isolated with silastic sheet. In the group I, all sides of flap were repaired with blockage of silastic sheet insertion. Three, two, and one sides of flap were treated with same method in the group II, III, and IV respectively. Other sides of flaps were repaired without blockage, so all sides of flap were repaired in the group V. At 14 days later, the survived rate of each flap was evaluated according to the numbers of the repair sites. Histological examination was done for the evaluation of new vessel development quantitatively. Overall survived rates were increased with the number of repaired sites, but the group V only showed increased survival rate up to more than fifty percentile of the flap size with a significant difference statistically. New vessels were also increased in proportion with the number of repaired sites, and the repair site more than two had significant effect on the increased number of new vessels. In conclusion, at least more than three-fourth of flap circumference should be repaired in order to increase flap survival effectively under the condition of bed isolation.