Knowledge of herbivory tolerance of dominant plant species and their responses to varying grazing intensity is required for sustainable grazing management in semiarid rangeland ecosystems. In a field experiment, we studied the effects of simulated grazing 0%, 30%, 50%, and 80% leaf removal (control, partial, intermediate, and heavy leaf removal) of three dominant perennial legumes: Astragalus cyclophyllon, Astragalus effuses, and Hedysarum criniferum in semiarid rangelands of southwestern Iran for 2 yr and tested 1) differential herbivory tolerance in terms of regrowth ability of belowground and aboveground biomass (AGB), fecundity, and nonstructural carbohydrate reserves; and 2) trade-off between forage quantity and quality—less defoliated plants produce larger amount of biomass with less forage quality and vice versa. The results showed a reduction in AGB and reproduction in each of the three plants. However, three forbs exhibited some sort of variations in shoot and root growth, reproduction and forage quality under intermediate and heavy leaf removal. Leaf removal in individuals of A. effusus and H. criniferum are more likely to compensate for tissue loss by allocating the available resources to the expansion of root biomass, at the cost of AGB and reproductive effort. Leaf removal also led to an increase in forage quality in each of the three forb species, depending on the level of height removal. The control plant had a higher amount of biomass production and lower amount of crude protein compared with heavy leaf removal treatments (i.e., 80%), by contrasting responses of acid detergent fiber and neutral detergent fiber, indicating a trade-off between forage quality and forage quantity. The results indicate that there is a balance among forage quantity, quality, and regrowth ability of belowground and aboveground tissues in intermediate leaf removal treatments of these forb species. This balance is used to develop principles for grazing management of steppe rangelands because such a regime meets different criteria including forage quantity, quality, and nonstructural carbohydrate reserves.