The geoid minima in the Indian Ocean and North Eurasia are separated by the Tibetan Geoid Ridge (TGR), yet the origin of TGR remains poorly constrained. Spherical harmonic analysis and geoid kernels indicate that the TGR has wavelengths of degrees 7-10 and is generated by density anomalies of degrees 7-10 in the mantle. By employing numerical geoid modelling with four different tomography-derived density structures, we determined that abundant high-density anomalies in the mantle transition zone beneath Tibet are responsible for TGR. Additionally, two previously proposed alternative evolving scenarios of the India-Tibet collision — Indian lithosphere subduction and Tibetan lithosphere dripping — are evaluated through geoid calculation. The former suggests abundant high-density structures in the Tibetan transition zone, which can generate a well-constrained TGR, while the latter does not. Therefore, we regard the Indian lithosphere subduction as a more plausible evolving scenario.