This paper studies job assignments for automated guided in container terminal settings under various conditions of uncertainty. An introduction to their operation is provided, along with a flexible dispatching algorithm, suitable for real-time control of AGVs. Using these concepts a new AGV dispatching approach is developed, capable of operating under uncertain conditions within a detailed container terminal model. Several performance indicators are presented, focused on generic features of vehicle operations as well the assessment of uncertainty levels inside the terminal. From the results of the simulation experiments, it is found that the proposed technique outperforms well known heuristics and alternative algorithms.