Abstract: The anisotropy of the structure and properties caused by the strong epitaxial growth of grains during laser powder bed fusion (L-PBF) significantly affects the mechanical performance of Inconel 718 alloy components such as turbine disks. The defects (lack-of-fusion, LoF) in components processed via L-PBF are detrimental to the strength of the alloy. The purpose of this study is to investigate the effect of laser scanning parameters on the epitaxial grain growth and LoF formation in order to obtain the parameter space in which the microstructure is refined and LoF defect is suppressed. The temperature field of the molten pool and the epitaxial grain growth are simulated using a multiscale model combining the finite element method with the phase-field method. The LoF model is proposed to predict the formation of LoF defects resulting from insufficient melting during L-PBF. Defect mitigation and grain-structure control during L-PBF can be realized simultaneously in the model. The simulation shows the input laser energy density for the as-deposited structure with fine grains and without LoF defects varied from 55.0–62.5 J·mm^–3 when the interlayer rotation angle was 0°–90°. The optimized process parameters (laser power of 280 W, scanning speed of 1160 mm·s^–1, and rotation angle of 67°) were computationally screened. In these conditions, the average grain size was 7.0 μm, and the ultimate tensile strength and yield strength at room temperature were (1111 ± 3) MPa and (820 ± 7) MPa, respectively, which is 8.8% and 10.5% higher than those of reported. The results indicating the proposed multiscale computational approach for predicting grain growth and LoF defects could allow simultaneous grain-structure control and defect mitigation during L-PBF.