Nanophotonic systems facilitate a far-reaching control over the propagation of light and its interaction with matter. In view of the increasing sophistication of fabrication methods and characterisation tools, quantitative computational approaches are thus faced with a number of challenges. This includes dealing with the strong optical response of individual nanostructures and the multi-scattering processes associated with arrays of such elements. Both of these aspects may lead to significant modifications of light-matter interactions. This article reviews the state of the recently developed discontinuous Galerkin finite element method for the efficient numerical treatment of nanophotonic systems. This approach combines the accurate and flexible spatial discretisation of classical finite elements with efficient time stepping capabilities. The underlying principles of the discontinuous Galerkin technique and its application to the simulation of complex nanophotonic structures are described in detail. In addition, formulations for both time-and frequency-domain solvers are provided and specific advantages and limitations of the technique are discussed. The potential of the discontinuous Galerkin approach is illustrated by modelling and simulating several experimentally relevant systems.