Global Navigation Satellite System (GNSS) multipath channel models are fundamental and critical for signal simulation and receiver performance evaluation. They also aid the designing of suitable multipath error mitigation algorithms when the properties of multipath channel are available. However, there is insufficient existing research on BeiDou Navigation Satellite System (BDS) signal multipath channel models. In this study, multipath channel statistical models are established on the basis of extensive datasets of the BDS B1I signal. A multipath parameter estimation algorithm is designed to extract information of multipath rays from the intermediate frequency data. The delay, power loss, Doppler fading frequency, and lifetime distribution models for static and dynamic vehicle platforms are established and compared, and the effects of the satellite orbit type and platform speed on the models are analyzed. The results reveal the detailed distribution and variation characteristics of the multipath parameters and are valuable for the development of accurate urban navigation systems.