This paper studies the redeployment scheme for wireless sensor networks when the sensor node is limited mobility and proposes a genetic algorithm based redeployment scheme (GARS) when the motion model for sensor node is a flip. The gene of GARS is the jump direction of sensor node. The fitness function of GARS is taken into account the coverage maximize and the total moved distance minimize. Simulation results show that this scheme is better than the FBSD for isotropic sensing model and it is effective to improve network coverage for directed sensing model. The scheme achieves the optimal motion planning for sensor node in the isotropic sensing model.