In previous papers (Colotta, V. et al. Arch. Pharm. Pharm. Med. Chem. 1999, 332, 39. Colotta, V. et al. J. Med. Chem. 2000, 43, 1158) we reported the synthesis and binding affinity at bovine (b) A(1) and A(2A) and human (h) A(3) adenosine receptors (ARs) of the 4-amino-6-benzylamino-2-phenyl-1,2,4-triazolo[4,3-a]quinoxalin-1-one (compound A) which resulted in a potent and selective A(2A) AR antagonist. Compound A provided the lead compound of a series of 6- or 8-(hetero)arylalkylamino-4-amino-2-phenyl-1,2,4-triazolo[4,3-a]quinoxalin-1-one derivatives (compounds 1-20) which are the object of this paper. Most of the newly synthesized compounds are inactive at hA(3) ARs while they possess both nanomolar bA(2A) affinities and different degrees of bA(2A) versus bA(1) selectivity. The binding data show that hydrophilic substituents on the benzyl moiety are the most profitable for bA(2A) receptor affinity. Furthermore, their steric hindrance seems to play an important role for the bA(2A) AR interaction, thus suggesting that the 6-aralkylamino moiety of these ligands interacts with a size-limited binding pocket of this AR subtype. Thus, the SAR studies provided us some new insights about the structural requirements of the bA(2A) AR recognition site.