Tolerance is defined as the ability of one cultivar to yield more than another cultivar under similar disease severity. If both cultivars suffer an equal loss in healthy (green) leaf area duration (HAD) over the grain filling period due to disease presence, then the yield loss per unit HAD loss is smaller for a more tolerant cultivar. Little is understood of what physiological and developmental traits of cultivars determine disease tolerance. In this study, we use a mathematical model of wheat to investigate the effect of a wide range of wheat phenotypes on tolerance. During the phase from stem extension to anthesis, the model calculates the assimilate source and sink potential, allowing for dynamic changes to the source-sink balance by partitioning assimilates between ear development and storage of water-soluble carbon (WSC) reserves, according to assimilate availability. To quantify tolerance, rates of epidemic progress were varied on each phenotype, leading to different levels of HAD loss during the postanthesis, grain-filling period. Model outputs show that the main determinant of tolerance is the total amount of assimilate produced per grain during the rapid grain-fill period, leading to a strong positive correlation between HAD per grain and tolerance. Reductions in traits that affect carbon assimilation rate and increases in traits that determine the amount of structural biomass in the plant increase disease tolerance through their associated reduction in number of grains per ear. Some of the most influential traits are the canopy green area index, carbon use efficiency, and leaf specific weight. Increased WSC accumulation can either increase or decrease tolerance. Furthermore, a cultivar is shown to be maximally tolerant when a crop is able to just fill its total sink size in the presence of disease. The model has identified influential functional traits and established that their associations with tolerance have a mechanistic basis.