Transition metal nitride-based thin films are commonly used as protective coatings in applications with demanding environments such as cutting tools or turbine blades. The high demand for such coatings drives a significant research effort to improve their hardness, fracture toughness, thermal stability and resistance against wear, oxidation and corrosion. One approach for improvement is to carefully introduce defects in the material systems to tune their properties. Another possibility is to design a microstructure, such as creating a multilayered system. In this work, atomisitc simulations within the scope of Density Functional Theory (DFT) and classical molecular dynamics (MD) were used to investigate the impact of microstructure and defects on the mechanical properties of cubic transition metal nitride thin films. Atomistic simulations allow us to study phenomena and processes at the atomic level granting access to information that is not accessible otherwise, or only with great efforts. Using DFT simulations in this thesis, we could show that vacancies in the MoAlN system stabilise the metastable cubic MoN and improve the elastic constants for low Al contents. The role of the microstructure, namely the superlattice architecture, was studied on AlN/CrN, TiN/CrN and AlN/TiN multilayers. The fracture toughness of AlN/CrN and TiN/CrN superlattices was predicted using DFT. The results show that a discrete interface description is necessary to model the systems and that approximations (e.g., the method of Grimsditsh and Nizzoli) can sometimes lead to wrong results. Our calculations also revealed that the magnetic nature of CrN is not reflected in lattice parameter oscillations. Additionally, we calculated the energy barrier while moving a twin boundary in CrN for various processes and confirmed that an asynchronous process is energetically preferred. It results in a termination of the twin boundary with a nitrogen plane, as observed experimentally. Next, we performed tensile loading and nanoindentation simulations using molecular dynamics for the AlN/TiN system with different bilayer periods. Utilising the larger number of atoms in MD calculations allowed us to study the development of dislocation networks, phase transformations and cracks. We could show toughening for small bilayer periods driven by a phase transformation in the AlN and differences in the failure mechanisms dependent on the loading direction and bilayer period. The nanoindentation simulations pointed towards intermixing of the layers, driven by the plastic deformation; this leads to weakening the superlattice structure.