The design of pharmaceutical tablets involves the determination of formulation parameters that define the tablet composition and internal microstructure. These parameters must be chosen so that the release of an active pharmaceutical ingredient (API) from the tablet follows a prescribed dissolution curve. In the case of immediate release formulations, the dissolution process typically consists of tablet disintegration, followed by the dissolution of the disintegration fragments. In order to find the appropriate values of formulation parameters, numerous experiments are typically required. In the present work, we propose a computational methodology for in silico design of tablet formulations with the aim of reducing the amount experimental work required during tablet design. The methodology is based on the coupling of two modelling approaches: (i) tablet fragmentation triggered by the swelling of the disintegrant is simulated by the discrete element method (DEM), and (ii) the dissolution of the resulting population of disintegration fragments is simulated using a finite volume gridbased model. The final API release curve is then obtained by the superposition of dissolution curves originating from the individual disintegration fragments. Using directly compressed tablets containing ibuprofen as the API and croscarmellose sodium as the disintegrant, the model was validated against experimental data. The fragment size distribution was evaluated by static light scattering and the dissolution profiles were obtained by a standard dissolution apparatus with UV/Vis spectroscopic detection. We demonstrate that the computational methodology is able to quantitatively predict the effect of disintegrant content and API primary particle size on the fragment size distribution and the final dissolution profiles, and is therefore useful as a tool for computerassisted tablet formulation design. (C) 2020 Elsevier B.V. All rights reserved.