Modelling organic devices is an outstanding challenge because device performance is very sensitive to how the molecules are packed and the films are highly disordered. An understanding of charge and exciton (bound electron-hole pair) transport in these materials is important if organic light-emitting diodes are to be exploited in displays, lighting, photovoltaics, transistors, and sensors. This paper discusses methods we have pioneered for predicting charge and exciton transport, in which polymer chains are explicitly modeled and charge and exciton transfer rates are taken from electronic structure theory. Monte Carlo and drift diffusion device models that link device performance with morphology are also covered. The focus here is on polymers, but there is much in common with small molecule organic materials.