Weed seeds initiate most weed invasions of arable fields, yet there is relatively little information on the value of managing weed seed banks. Matrix population models were used to examine the relative importance of managing weed seed banks, in relation to other life stages, for four model weed species with varying life histories. Simulations for giant foxtail and common lambsquarters, summer annual weeds of arable fields; garlic mustard, an obligate biennial invasive weed of temperate forests; and Canada thistle, a perennial weed of pastures and arable fields, were run under conditions of varying population density and efficacy of seedling control. The models were subjected to elasticity analysis to determine what happened to weed populations when different life stages were targeted. Losses from the dormant seed bank were most important for summer annual weeds, of intermediate importance for biennial weeds, and of low importance for perennial weeds. More effort is needed to develop weed seed-bank management techniques for summer annual weed species as part of integrated weed management systems.