In this paper, we examine how to develop realistic-looking handwritten letters and numbers using Generative Adversarial Networks (GANs). To produce clear and visually compelling outputs, we suggest a model based on WGAN-GP, a stable variant of GANs. Our technology can produce particular characters or numbers as required by enabling user-defined character inputs. This study not only shows a high model performance, but also suggests real-world applications in digital document design, education, and automated writing systems.