An improved algorithm is discussed in the paper which is combined the Chan algorithm with the Taylor series expansion method to deal with the problem that the Chan algorithm has big errors in the NLOS environment and the Taylor series expansion method requires an estimate of the location as the initial value. The influence of NLOS is removed gradually by bringing the stretch factor into the Chan algorithm. And then the positioning results are brought into the Taylor algorithm. The simulation results show that this improved algorithm can effectively restrain the influence of NLOS.