Abstract: In order to study the possibility of amaranth cultivation in arid and semi arid regions of Iran, in respect to yield production, a case study was conducted at the Abourayhan campus (a branch of Tehran University); located at Pakdasht area of Varamin as an arid and semi-arid area. This experiment was done on 16 amaranth genotypes in growing season of 2001-2002 to evaluate yield, yield components and protein contents of grains. Present results indicated that grain yield varied from 3.445 to 2.265 t ha-1 for Amaranthus cruentus, while these values varied from 1.385 to 1.945 t ha-1 for Amaranthus hypochondriacus (mean values were 2.95 and 1.65 t ha-1 for A. cruentus and A. hypochondriacus, respectively). Putting these results with obtained yields showed an average yield of 2000 kg ha-1. Higher mean 1000 grain weight (68.47 versus 58 g), harvest index (55.16 versus 33.4%) and total dry weight percentage (46.9 versus 38.2%) for A. cruentus resulted in higher grain yield than A. hypochondriacus, while A. hypochondriacus genotypes showed higher protein content than A. cruentus (mean values of 15.7 and 14.43%, respectively). it concluded that amaranth species (especially A. cruentus spp.) could be cultivated successfully in arid and semi arid regions, if we select appropriate genotypes and the best agronomic options to achieve higher yields. This crop can be investigated as a new crop to extend arable marginal lands and sustaining the agricultural systems of stressful environments.