A multihop localisation algorithm that incorporates the distance estimation bias for 3D wireless sensor networks is presented. This work is based on the discovery that the multihop distance estimation errors approximately obey normal distribution with various biases. First, the normal node feasible regions are obtained by calculating the intersections or pseudo-intersections of bounding cubes. Then, the node positions and the distance estimation bias are solved simultaneously by using the projected Levenberg-Marquardt method. Simulation results confirm the effectiveness of this algorithm.