The accuracy of RSSIbased ranging is severely affected by multipath effect and external radio interference in underground mine tunnel environments. To solve this problem, a novel RSSIbased ranging algorithm is proposed which overcomes the multipath effect by applying geometrical optics rules in analyzing RSSI measurements in both Lineof Sight and NonLineofSight scenario. The proposed algorithm can reconstruct a function for path loss and the physical distance which can help minimize the ranging error caused by multipath transport. Extensive real experiments are conducted in underground tunnel environments and the accuracy of the proposed ranging algorithm is analyzed. The impact of representative system parameters on the ranging, including multipath attenuation factor and path loss exponent, are discussed. Experiment and simulation results demonstrate that the proposed algorithm outperforms conventional RSSIbased ranging method without using any radio map information nor a calibration stage.