<p>Industrial production of high-quality components by laser powder-bed fusion (PBF-L) additive manufacturing requires careful assessment of the combined role of part design, material properties, and process parameters in determining the final component’s mechanical properties. Various uncertainties arising throughout the PBF-L process affect the quality of the printed components. Therefore, understanding the impact of different uncertainties on part quality is crucial. In this work, we present a multiscale framework for the quantification and propagation of uncertainty across the PBF-L process. Here uncertainty in material property is considered, where thermal conductivity, specific heat, and density are considered random, and the effect of material uncertainty is studied at the scale of the powder and the scale of the part. The variation in material properties in part is modeled as Gaussian random fields, and the spatial variation of mechanical properties of the printed part is quantified. Also, we quantitatively show how uncertainty propagates from the scale of the part to the scale of the powder, and then back to the scale of the part. We demonstrate the framework with the case of two coupon bars, one printed vertically and the other horizontally on the build plate. We show that, because of the larger impact of heat loss through the baseplate, uncertainty is reduced in horizontal bars as compared to vertical bars, where overheating is possible. According to our research, printing parts having the most contact possible with the baseplate can effectively lower part characteristic uncertainties and improve part quality in real-world additive manufacturing applications.</p>