We study the effect of spatially-correlated heterogeneity on isothermal drying of porous media. We combine a minimal pore-scale model with microfluidic experiments with the same pore geometry. Our simulated drying behavior compares favorably with experiments, considering the large sensitivity of the emergent behavior to the uncertainty associated with even small manufacturing errors. We show that increasing the correlation length in particle sizes promotes preferential drying of clusters of large pores, prolonging liquid connectivity and surface wetness and thus higher drying rates for longer periods. Our findings improve our quantitative understanding of how pore-scale heterogeneity impacts drying, which plays a role in a wide range of processes ranging from fuel cells to curing of paints and cements to global budgets of energy, water and solutes in soils. Plain Language Summary Drying of porous media such as soils, cement, food or fuel cells is an important process in many natural and industrial systems. Drying in soils is of particular environmental importance, as it controls the transfer of water, energy and solutes between the subsurface and the atmosphere. We study the effect of spatial correlation in particle sizes on drying rate and extent, by combining pore-scale modeling with state-of-the-art microfluidic experiments of the same pore geometry. Our simulations compare favorably with experiments, considering the large sensitivity of the emergent behavior to the uncertainty associated with manufacturing errors. We show that increasing the correlation length in particle sizes-the distance in which the probability to encounter particles of similar sizes is high-promotes preferential invasion of clusters of large pores, prolonging liquid connectivity and surface wetness and thus higher drying rates for longer periods. Our findings improve our understanding of how pore-scale heterogeneity, inevitable in most porous materials, affects their drying.