From the viewpoint of materials science, the vast majority of biological materials are organic–inorganic composites with a hierarchical organization spanning over multiple levels from the molecular scale to the macro‐scale. What makes them interesting as materials is that they have been optimized during evolution to perform vital functions within the specific eco‐physiological constraints imposed on living organisms. These functions are very diverse and can be e.g. of mechanical, locomotive, optical or sensory nature, and frequently combinations of them. The required diversity of physical properties is caused by structural and chemical alterations at different hierarchical levels utilizing the morphological and genetic prerequisites available to the organism. In order to understand the design principles of such materials with specific functions, it is necessary to study the relationship between their structure, composition and the resulting physical properties. This is usually done using an experimental approach, where materials science offers a large variety of methods to study microstructure, chemical composition and mechanical properties and behaviour. In practice, however, it is frequently not possible to establish and validate the overall structure–property relationships for a biological material owing to the complex structural hierarchy and methodological constraints. Numerical multi‐scale models are very elegant and versatile tools to overcome these inherent shortcomings since they can systematically describe materials properties from the atomic up to the macroscopic scale.