Single image super-resolution (SISR) is a fundamental task in computer vision that aims to generate high-resolution images from low-resolution inputs. The Super-Resolution Convolutional Neural Network (SRCNN) is a widely used method for SISR, but it has limitations in capturing complex image structures. In this paper, we propose a transformation of the SRCNN model using complex-valued neural networks to address these limitations. Complex-valued neural networks have the potential to capture both magnitude and phase information, which can lead to improved reconstruction quality. We present a detailed methodology for incorporating complex-valued operations into the SRCNN model and evaluate its performance using various evaluation metrics. Experimental results demonstrate the superiority of the complex-valued network over the traditional SRCNN, highlighting the potential advantages of complex-valued neural networks in enhancing single image super-resolution.