This paper deals with research into advanced hierarchical graph structures which have been developed for knowledge representation in computer-aided design and robotics. Complex real-life engineering problems still need useful graph-based models. A new approach to represent models preserving the intrinsic nature of problem structures using graphs with many hierarchies is proposed. Different types of hierarchical dependencies (geometrical, functional, etc.) between design parts can be expressed in one structure. Moreover, the possibility of selecting only problem-related hierarchies reduces the number of data to be analysed in the process of reasoning about considered designs. This approach is illustrated by applying multi-hierarchical graph structures to represent complex data used for robot control algorithms.