The central concern of this work is to convince the reader, how the exponential availability of high-performance computing power, will open another complementary branch concerning the selection, understanding and improvement of novel nanostructured functional materials. This raises the expectation that the computational design and characterization of functional materials will become one essential step in this quest -- on par with experimental techniques, like high-throughput characterization on combinatorial materials libraries, transmission electron microscopy or synchrotron based x-ray analysis methods. Their combination can yield a complete picture ranging from the electronic structure of an individual atom to the macroscopic functional properties, which will provide us with what we call understanding and - derived from this - heuristic guidelines for the development of better materials. The strong benefit from electronic structure theory is demonstrated at three examples of nanostructured materials, which are characterized by different degrees of complexity. These are elemental nanoclusters (in particular iron), binary nanoparticles for data recording purposes and bulk-like magnetic shape memory alloys, which form a new class of magnetomechanical actuator materials. In particular, large scale density functional theory calculations corroborate that close-packed Fe nanoclusters are instable against a partial transformation along the so-called Mackay-path, which is a continuous, diffusion-less transformation pathway between a cuboctahedral and icosahedral both arrangement. Intermediate stages of this transformation exhibit a bcc-like coordination of the sub-surface atoms which stabilizes this structure against the pure close packed morphologies. Stoichiometric Fe-Pt nanoparticles are a widely discussed as a candidate future data recording materials. However, the required hard magnetic properties is only obtained for a single crystalline layered ordering of the elements. The calculations in this work show that this should not be expected for particles of 4nm and smaller, which will then be characterized by a different structure and ordering. Furthermore, it is demonstrated that the contributions from surface, stoichiometry and disorder to the magnetocrystalline anisotropy will be significant as well. Finally, by inspection of two paradigmatic materials classes, the ordered Ni-Mn-Ga Heusler compounds and disordered fcc Fe-Pd alloys, we will discuss important criteria which define a successful magnetic shape memory system: These are a flat energy profile along the transformation pathway, which changes with temperature through an effective magnetoelastic coupling mechanism. In addition, a phonon anomaly which arises from electronic instabilities which introduces a periodic modulation and supports the shear mechanisms leading to the formation of periodic nano-twinned representations of the non-modulated ground state with a very low interface energy. These, in turn, lead to a hierarchical microstructure which allows for very mobile interfaces. Due to a sufficiently larger magnetocrystalline anisotropy of the structure these interfaces can be shifted by the very low energy supplied in a magnetic field leading to a large macroscopic strain.