Localization is a very important issue to wireless sensor networks. The sensor node requires accurate location information in order to achieve the purpose of real-time monitoring and transmission of information. The more the sensor nodes and GPS modules are used in localization, the higher the localization accuracy can achieve, but employing more sensor nodes also leads to high costs. Thus, how to improve the localization accuracy while solving the problem of high costs has garnered attention. In this paper, an area-restriction based localization (ARBL) algorithm is proposed. The ARBL algorithm uses both range-based and range-free schemes, which combines area restriction setting and the sampling points from Monte Carlo localization (MCL) to improve the localization accuracy. The simulation results show that the ARBL algorithm proposed in this paper outperforms DuRT, MMRL and RL algorithms that employ a mobile anchor node, regardless of the number of the sensor nodes and the moving speed of the anchor node.