Problem solving in complex situations or systems requires an appropriate and reliable understanding of the problem. We propose a framework that combines the Iceberg and Situational Awareness models. The goal is to raise decision makers' awareness so that, depending on the system layers, problems can be perceived as events and trends within the system structure. Depending on what has been perceived, decision makers are then able to understand the impact of problems or their underlying causes and conditions. The appropriateness of the perceived and integrated information depends on whether it helps decision makers answer their questions and whether it corresponds to reality. Reliability depends on the completeness and certainty of the information presented. The perception of system properties is influenced by the (in)transparency of the real system and the prevailing paradigm and mindset of the observer. So, we propose a system taxonomy within the framework. This summarizes system properties of simple, complicated and complex systems and the resulting behavioral characteristics identified in the literature. Decision makers can use the taxonomy to explore the properties of real systems to develop an appropriate system model and assess its reliability, or to evaluate existing system models whether, or to what extent, they are suitable for decision support. This is because, based on these models, decision makers develop a corresponding understanding of the problem, upon which their decisions for possible solutions are based.