Abstract: With the rapid development of artificial intelligence technology, small molecule generation models have emerged as a significant research direction in the field of drug discovery. These models, including Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and diffusion models, have proven to possess remarkable capabilities in optimizing drug properties and generating complex molecular structures. This article comprehensively analyzes the application of the aforementioned advanced technologies in the drug discovery process, demonstrating how they supplement and enhance traditional drug design methods. At the same time, it addresses the challenges facing current methods in terms of data quality, model complexity, computational cost, and generalization ability, with a prospect of future research directions.