Brachypodium, a monocot grass with wide distribution in temperate areas, has been considered a new model plant for many grass species, such as turfgrass or miscanthus, due to its small genome size, self-pollination, rapid life cycle, higher seed yield, and small stature. The objectives of this study were to compare the genetic diversity of natural and artificial populations, and to determine the effects of ${gamma}$-radiation on genetic variability when evaluated by amplified fragment length polymorphism (AFLP) markers. Two populations used in this study include 66 plant introductions (PI) from the U.S. Department of Agriculture (USDA; POP1), and 43 mutants derived from one of the plant introductions, Bd43 (PI 227011; POP2) treated with a ${gamma}$-irradiation ranging from 100 to 600 Gy. The highest performance of the phenotypic traits was observed at 200 Gy for germination rate, 100 Gy for height, and 600 Gy for tiller number, implying that the $LD_{50}$ for determining of optimum dosage depends on the physiological parameters measured. Based on AFLP analysis, POP2 showed higher polymorphism (79.4%), PIC (polymorphism information content; 0.162), and genetic diversity (0.195) than POP1 (59.4%, 0.113, and 0.130, respectively). Those three genetic parameters were evaluated for the mutants derived at various ${gamma}$-radiation dosages and were found to be the highest when irradiated at 300 Gy. Artificial mutagenesis using ${gamma}$-radiation improved genetic diversity compared to that of the natural population, and 300 Gy was a useful dosage to enlarge genetic variability in Brachypodium. Results suggest that a dosage maximizing the genetic diversity when evaluated with AFLP markers is an alternative index to $LD_{50}$ for determining the optimum level for mutation induction.