It is very important for wireless sensor network to position the nodes' location because location information is favorable for providing network service such as geographic routing, people tracking and so on. After researching the range-based atomic multilateration algorithm carefully, this paper presents a novel positioning algorithm based on Adaptive Genetic Algorithm (AGA). The new positioning algorithm uses adaptive genetic algorithm to search the optimal solution of nonlinear multivariate positioning equations set. Furthermore, the proposed algorithm uses a new model to estimate range error in order to improve the localization accuracy. Simulation results show that the proposed scheme has better robust performance.