China's Tianwen-1 successfully landed in the southern Utopian Planitia of the Martian surface on 15 May 2021. The Zhurong Rover, equipped with a high frequency full polarimetric Rover Penetrating Radar (RoPeR), travelled 1,921 m to investigate the shallow geological structure and material composition of the Martian weathered layer. In this study, we propose a new processing strategy to estimate surface relative permittivity using the HH and VV reflections of the high frequency RoPeR data. This new strategy is based on the induced field rotation effect, which occurs when orthogonally-polarised electromagnetic waves propagate into an uneven surface with incident angles. 3D time-domain finite-difference simulations were performed using random surfaces with various relative permittivities under the same geometry as the Zhurong Rover. Polarimetric alpha angle versus relative permittivity was then calculated based on the simulation results. At the same time, direct coupling removal, band-pass filtering and channel calibration were performed on the real RoPeR data and clear surface reflections were extracted. The surface reflection amplitudes of the HH and VV were then obtained and the polarimetric alpha angle calculated. Finally, relative permittivity was estimated through the relationship obtained from the simulation results. The average value of the relative permittivity estimated by the proposed approach is 3.292, with a standard deviation of 0.235. This result is consistent with that obtained by orbiting Radar Systems and the low frequency RoPeR system. This study will contribute to the further signal processing and accurate interpretation of real radar data captured by way of RoPeR on Mars.