Cicer milkvetch (Astragalus cicer L.), a long‐lived perennial forage legume, has many attributes that make it well suited for permanent pastures in a sustainable agricultural system. Its relatively poor seedling vigor, however, has reduced its potential. A selective, postemergence herbicide, e.g., 2,4‐D [(2,4‐dichlorophenoxy)acetic acid], could improve establishment by controlling broadleaf weeds. Therefore, this study was conducted to evaluate the use of phenotypic recurrent selection for increasing the tolerance of cicer milkvetch seedlings to 2,4‐D. Seedlings were treated in the greenhouse at the three‐leaf stage of development with a 1.0 kg a.i. ha− rate of 2,4‐D. Five cycles of recurrent selection for regrowth at the apical meristem following treatment with 2,4‐D increased extended seedling height (total length) 38% more than that of the cultivar Monarch (Cycle 0). Seedlings were visually selected for extended height about 4 wk after herbicide application. Selection for extended height reduced root and total seedling dry weights 21 and 13%, respectively, from the Cycle 0 to the Cycle 1 population. For Cycles 1 through 5, however, extended height was significantly correlated with shoot (r = 0.99), root (0.93) total seedling (0.97) dry weights, and apical meristem regrowth score (0.97), but not with crown meristem regrowth score ( − 0.50). Recurrent selection for increased regrowth at the apical meristem following application of 2,4‐D was an effective and inexpensive procedure for increasing the tolerance of citer milkvetch to 2,4‐D.