As a shared passenger-and-freight transportation mode, the metro-based underground logistics system (M-ULS) can significantly alleviate negative effects caused by ground transportation vehicles and improve urban sustainability. Against this background, this paper addresses dedicated freight train operation and freight assignment problems in the M-ULS without affecting passenger service during the off-peak period. More importantly, this paper proposes a flexible skip-stop strategy to reduce the increased operational cost of freight trains. To hedge against uncertainty, this paper formulates a distributionally robust chance-constrained programming model by limiting uncertain freight demands to a Wasserstein ambiguity set. Moreover, the distributionally robust optimization model for a Wasserstein ball with the 1-norm can be transformed into a mixed-integer linear program, which allows for an efficient solution via CPLEX, thereby enabling metro operators to use it. To implement this study in real-life applications, a case study is conducted based on the Beijing Subway Yizhuang Line. The computational results verify the superiority of the skip-stop strategy and distributionally robust optimization approach proposed in this study. Finally, some managerial suggestions are provided for the metro operators.