Hydrothermal synthesis based upon the use of Al³⁺ as the dopant and/or ethanol as the solvent is effective in promoting the growth of hematite into nanoplates rich in the (001) surface, which is highly active for a broad range of catalytic applications. However, the underpinning mechanism for the flattening of hematite crystals is still poorly comprehended. To close this knowledge gap, in this work, we have attempted intensive computational modelling to construct a binary phase diagram for Fe₂O₃–Al₂O₃ under typical hydrothermal conditions, as well as to quantify the surface energy of hematite crystal upon coverage with Al³⁺ and ethanol molecules. An innovative coupling of density functional theory calculation, cluster expansion and Monte Carlo simulations in analogy to machine learning and prediction was attempted. Upon successful validation by experimental observation, our simulation results suggest an optimum atomic dispersion of Al³⁺ within hematite in cases when its concentration is below 4 at% otherwise phase separation occurs, and discrete Al₂O₃ nano-clusters can be preferentially formed. Computations also revealed that the adsorption of ethanol molecules alone can reduce the specific surface energy of the hematite (001) surface from 1.33 to 0.31 J m⁻². The segregation of Al³⁺ on the (001) surface can further reduce the specific surface energy to 0.18 J m⁻². Consequently, the (001) surface growth is inhibited, and it becomes dominant after the disappearance of other surfaces upon their continual growth. This work provides atomistic insights into the synergistic effect between the aluminium textural promoter and the ethanol capping agent in determining the morphology of hematite nanoparticles. The established computation approach also applies to other oxide-based catalysts in controlling their surface growth and morphology, which are critical for their catalytic applications.