This paper proposes a new method for service restoration of distribution network with the support of transportable power sources (TPSs) and repair crews (RCs). Firstly, a coupling model of distribution networks and vehicle routing of TPSs and RCs is proposed, where the TPSs serve as emergency power supply sources, and the RCs are used to repair the faulted lines. Considering the uncertainty of traffic congestion, the probability distribution of the travel time spent on each road is derived based on the Nesterov user equilibrium model, and a two-stage stochastic program is formulated to determine the optimal routings of TPSs and RCs. To efficiently solve the proposed stochastic mixed-integer linear program (MILP), a two-phase scenario reduction method is then developed to scale down the problem size, and an adaptive progressive hedging algorithm is used for an efficient solution. The effectiveness of the proposed methods and algorithms has been illustrated in a modified IEEE 33-bus system.