Abstract: In some cases such as sudden satellite failure, bad weather environment, and deterioration of satellite and ground channel environment, the positioning effect of global navigation satellite system (GNSS) becomes poor or even invalid, leading to the inaccurate absolute position of remote sensing platform, and seriously affecting the performance of remote sensing mapping. Therefore, a three-dimensional relative positioning algorithm combining phase ranging and adaptive Levenberg-Marquardt (LM) method is proposed in this paper. The algorithm involves two steps. In the coarse positioning, the phase ranging method of Chirp signal was proposed to enhance the ranging accuracy and improve the coarse positioning accuracy under the framework of Least Squares algorithm. In the fine positioning, an adaptive LM algorithm was proposed to improve the convergence effect and positioning accuracy of the traditional LM algorithm. Simulation results show that compared with the traditional methods, the three-dimensional positioning accuracy of the remote sensing platform relative to anchor nodes by the proposed method has been improved by two orders of magnitude. Since the absolute position of anchor nodes is known, the proposed method can further provide reliable absolute position information for the remote sensing platform in a GNSS-denied environment, and ensure the mapping quality of the remote sensing platform.