The day-to-day developments in computer-aided drug design methods have improved the identification process of potential drugs. The de novo design creates novel chemical entities to increase the search space of potential drug-like candidates. Evolutionary-based algorithms and deep learning models like Generative Adversarial Networks (GANs) are extensively used in the de novo design. In this work, we are studying two models, a Genetic algorithm (GA) based on the evolutionary algorithm and WGAN-GP, which is an improved GAN model. The generated molecules have been further evaluated using drug-likeness, synthesizability, solubility and validity measures. It has been found that both models produced novel drug-like molecules and obtained better or similar scores compared to other baseline models.