To address the issue of ensuring reliable power supply to critical loads in urban distribution networks during extreme ice disasters, a pre-disaster mobile energy storage pre-deployment strategy considering time-varying faults in distribution networks is proposed to enhance distribution network resilience. First, combining the meteorological forecast data and the geographical information of the distribution network, time-varying fault probabilities of distribution lines are calculated based on the spatiotemporal propagation characteristics of ice disasters. A set of time-varying distribution network fault scenarios are then generated using time-sequential Monte Carlo simulation. Next, considering the requirements of pre-deployment of mobile energy storage for reliable power consumption during ice disaster scenarios, a stochastic optimization model is constructed to determine the optimal pre-disaster deployment of mobile energy storage systems, to minimize the load losses caused by dynamic adjustment of mobile energy storage connection points under various fault scenarios. Finally, case studies using an improved IEEE33-bus distribution system are conducted to verify the effectiveness of the proposed strategy.