Coal is highly heterogeneous at multiple scales, such that the characterisation of coal lags behind that of conventional reservoir rocks. This paper demonstrates the capability of a digital coal model that can be used to characterise the multiscale structure of coal as well as its petrophysical properties at the core scale (cm). X-ray micro-computed tomography (micro-CT) is applied to obtain the internal cleat structure of a heterogeneous fractured coal sample, based on which statistics of coal lithotype distributions and cleat geometrical properties are extracted. Digital coal models are constructed stochastically according to the measured statistics. Resulting models preserve the core scale (cm) heterogeneity and pore scale (mm to mm) cleat properties of the imaged coal sample. Furthermore, petrophysical evaluation is performed based on the stochastic digital coal models. We find that the generated digital coal models can provide permeability estimation with errors of 26.7% in comparison to that of the original micro-CT data. Since the presented model is able to preserve the multiscale heterogeneities as well as petrophysical properties of coal, our work provides an alternative to segmented micro-CT images. Therefore, we can avoid the challenges that are inherent to micro-CT images, such as segmentation errors, size and resolution limitations. The digital coal model can also be a tool of linking the rock microstructure to petrophysical properties at the core scale, such that we can predict the permeability of coal based on its geometrical measures rather than expensive laboratory experiments. (C) 2017 Elsevier Ltd. All rights reserved.