Online phishing attacks are becoming increasingly sophisticated, posing significant challenges to cybersecurity systems. Existing solutions, such as blacklist/whitelist-based URL filtering and traditional machine learning approaches, remain limited in their adaptability to new attack patterns. This study proposes a novel phishing domain detection method that integrates data augmentation using various Generative Adversarial Networks (Basic GAN, SeqGAN) with machine learning classifiers. Using GANs to generate synthetic phishing domain samples, our approach enhances the diversity and quality of training data, thereby improving the detection of previously unseen attack patterns. Experiments on real-world datasets demonstrate that the proposed method achieves an accuracy of 99.3%, outperforming conventional solutions. These results highlight the potential of GAN-based data augmentation to strengthen phishing detection systems significantly and offer a promising direction for future cybersecurity solutions.