A multi-scale, multi-physics modelling framework of selective laser melting (SLM) in the nickel-based superalloy IN718 is presented. Representative powder-bed particle distribution is simulated using the measured size distribution from experiment. Thermal fluid dynamics calculations are then used to predict melting behaviour, sub-surface morphology, and porosity development during a single pass scanning of the SLM process. The results suggest that the pores and uneven surface structure are exacerbated by increasing powder layer thicknesses. Predicted porosity volume fraction is up to 12% of the single track when 5 statistical powder distributions are simulated for each powder layer thickness. Processing-induced microstructure is predicted by linking cellular automatons – finite element calculations indicate further that the cooling rate is about 4400 o C/s and grain growth strongly follows the thermal gradient giving rise to a columnar grain morphology if homogeneous nucleation is assumed. Random texture is likely for as-fabricated SLM single pass with approximately 8 Pm and 6 Pm grain size for 20 Pm and 100 Pm powder layer thickness fabrication. Use has been made of the cooling history to predict more detailed microstructure using a γ" precipitation model. With the short time scale of solidification and rapid cooling, it becomes less likely that γ" precipitation will be observed in the condition investigated unless a prolonged hold at temperature is carried out. Future work on extension of the proposed multiscale modelling approach on microstructure predictions in SLM to mechanical properties will be discussed.