In problem-based learning environments the learners work actively on problem tasks in order to learn specific subject-matter. Often the learning activities are applied in a collaborative way. Because it is impossible to train all variations of all tasks, the processes underlying effective performance should also be focused on. By means of guided group discussion and reflecting on the problems, the fostering of learning- to-learn skills is stimulated: a deep understanding of the performance will increase the probability that learners will perform well in situations not encountered previously. Guided by a coach the learners discuss about the problem tasks, and exchange experiences with each other. Although the current problembased learning programs focus on mastering individual skills, this could be extended to team skills as well. Team training programs are primarily aimed at the behavioral and cognitive requirements of team task performance. In many cases, technologically advanced learning environments are employed, like simulations, (distributed interactive) simulators, and virtual reality. In these learning environments complex problems can be practiced, requiring the team to work together in a coordinate way. Important conditions for effective team training, just as in individual problem-based learning, are adequate training scenarios, appropriate guidance by a coach, and ample opportunities for reflecting on the learning tasks. The paper focuses on the iterative process of designing problem-based team training scenarios. A set of guidelines will be described comprising the following categories: (A) general approach, (B) structure of a scenario, (C) contents of a scenario, (D) training strategy, (E) team performance and feedback, and (F) role of the instructor/observer. Our experiences with these guidelines in specifying Training Support Packages for training teams in networked simulators will be discussed as well.