Due to rising volatility, uncertainty, complexity, and ambiguity in- and outside the company boundaries, manufacturers of complex technical systems are confronted with unprecedented challenges affecting their development environment. Many projects fail to meet their economic or technical objectives, and established companies are rethinking their traditional development approaches which are oriented towards plannable stage-gate structures. Rather than orchestrating the entire project in detail and determining stakeholder requirements upfront, companies tend to establish agile development approaches that are characterized by the iterative development of prototypes as functional system increments. When applying agile techniques to the development of complex technical systems, however, existing frameworks known from the software industry do not adequately consider the complexity caused by strong interdependencies between and within the system elements. For this reason, the paper presents a model that structures the process of dependency-oriented prototyping, not only when selecting the next prototype to develop, but also when implementing the prototype in the upcoming development iteration. By forcing the early reduction of development-related uncertainty on the one hand, as well as minimizing and controlling the early determination of interdependent system elements on the other hand, market and technology knowledge can be generated quickly and the elaborated system design remains flexible in the course of the development project.