Generative AI has emerged as a powerful tool in healthcare, enabling the creation of high-quality synthetic datasets that mirror the complexity of real-world patient data while addressing privacy concerns. Personalized medicine, which tailors treatment plans based on individual patient characteristics, requires large datasets of patient demographics, genetic markers, drug attributes, and clinical outcomes to train predictive models effectively. However, accessing real-world medical data is often challenging due to privacy regulations and ethical concerns, creating a significant barrier to developing personalized treatment strategies. In this study, we developed a Generative AI-driven simulation system to generate synthetic datasets for the prediction of treatment efficacy in personalized medicine. The system produced 1000 patient records with 1075 features, including patient demographics, genetic markers, and clinical outcomes. After preprocessing the data, we applied machine learning models such as Logistic Regression, Random Forest, Support Vector Machines (SVM), and Neural Networks to predict treatment efficacy. Clustering techniques and association rule mining were also used to uncover hidden patterns in patient characteristics and drug responses. The results showed that Neural Networks and Random Forest were the most effective models, with accuracies of 97% and 96%, respectively. Key genetic features, particularly SNP_599 and SNP_163, were identified as the most influential predictors of treatment efficacy. Clustering analysis revealed distinct patient subgroups, providing actionable insights for personalized treatment approaches. The use of synthetic data alleviated concerns regarding patient privacy, demonstrating that Generative AI can be a viable solution for generating large-scale healthcare data for model training. The implications of this research are significant for personalized medicine. By using Generative AI to simulate patient data, researchers can bypass the limitations of real-world data availability while maintaining the statistical complexity required for accurate model training. The high performance of machine learning models in this study suggests that integrating AI-driven predictive systems into clinical workflows could enhance decision-making and lead to better patient outcomes. Future work will focus on refining the simulation system and integrating real-world data to further improve the generalizability of the models.