This paper proposes a Petri net (PN) representation of the activities performed by the key human operators for unloading/loading containers in an intermodal maritime container terminal (CT) with a low level of automation. These processes are the core of the export, import, and transshipment cycles executed in the terminal. The aim of this paper is to consider both the human component and the material handling resources, e.g., cranes and transporters, by defining an accurate model, which describes how to coordinate humans and use the system resources necessary for serving mother or feeder ships. The developed generalized stochastic PN-based model is of limited complexity and represents a complete, unambiguous, and readable model of the target process before coding it in the target simulation tool. The modular integrated model is tested and validated by the simulation of typical and perturbed scenarios of the Taranto CT, a real terminal that is taken as a case study for its complexity and similarity to CTs with multiple transport modes.Note to Practitioners-This paper proposes a new perspective for modeling, simulating, and controlling container terminals (CTs). Usually, the participation of human operators to the unloading/loading processes is neglected. However, a bad behavior of human operator can significantly modify the CT performances. Moreover, a few CTs are characterized by a low or partial level of automation. Then, the cooperation between the humans and the material handling equipment is important for executing several key tasks with efficiency (with respect to the deadlines), safety, and sustainability for humans. The complexity arising from a human-in-the-loop system demands for more accurate models that can reduce prediction errors and avoid unnecessary extra-equipment costs. Therefore, we propose an approach to explicitly consider the roles and main functions performed by a typical team of specialized operators. To verify the model accuracy and validate it, we simulate both standard and perturbed conditions of the Taranto CT. Given that the operators are usually assigned well-established tasks, the approach can be useful: 1) to improve the employment of humans and their supervisory activity; 2) to identify critical or faulty situations that involve operators or are caused by their errors and anomalous behaviors, thereby to improve safety and protect the handling equipment (e.g., by identifying delays and message losses in Manuscript operators' communication); and 3) to increase automation, by reassigning tasks from humans to automatic equipment.