Single Image Super-Resolution (SISR) is a very active research field. This paper addresses SISR by using a GAN-based approach with dual discriminators and incorporating it with an attention mechanism. The experimental results show that GDCA can generate sharper and high pleasing images compare to other conventional methods.