Recent advancements in image processing have highlighted the potential of combining various generative techniques for tasks like style transfer. This paper introduces a novel hybrid model that leverages the strengths of generative adversarial networks (GANs) and variational autoencoders (VAEs) to improve the quality and stability of style transfer. Our proposed GAN-VAE hybrid focuses on transferring artistic styles between images while maintaining structural coherence and producing high-quality, diverse results. By utilizing the adversarial learning framework of GANs alongside the latent space representation of VAEs, the model effectively captures both the content and stylistic features necessary for complex style transformations. Through experiments on several benchmark datasets, we demonstrate that our model surpasses standalone GAN and VAE-based approaches in terms of style fidelity, content preservation, and training stability. This hybrid approach opens new avenues for refining style transfer techniques, enhancing the visual realism and diversity of transferred styles.