Dynamic asset repositioning models are used in many freight transportation contexts to enable operators to cost-effectively serve demand which may be both geographically and temporally imbalanced. In this research, we develop a new robust approach for dynamic asset repositioning problems with uncertainty. Specifically, we model future demand uncertainty using intervals about nominal values. Further we limit the maximum total deviation from the nominal values via a parameter which controls the conservatism of the solution. We test these ideas on a weekly repositioning decision problem faced by global tank container operators. In this talk, the robust asset repositioning decision problem will be introduced, and necessary and sufficient conditions for robust flows will be developed for several important special cases relevant for tank container management.