In modern rail-road container terminals, container transshipment between trains and trucks is carried out by large gantry cranes that span over rail tracks and road lanes. Prior optimization approaches have mainly focused on the short-term scheduling of these crucial transshipment resources. In this work, we focus on a midterm planning level that assigns inbound trains to yard tracks such that trains can be serviced by yard resources on schedule. This planning task is especially crucial in larger terminal settings which comprise more than a single yard, which is increasingly typical in growing intermodal networks. We describe the planning situation we encountered at a large rail-road terminal in Germany and propose a mixed-integer model formulation that captures the core of the decision problem. Further, we carry out a complexity analysis of problem variants and develop a heuristic solution procedure in a simulated annealing framework based on these structural insights. The approach is tested in a comprehensive computational study which demonstrates the viability of the approach to solve instances of real-world size.