With the rapid advancement of emerging technologies such as cloud computing and the Internet of Things (IoT), the world is entering a new era of data-driven economy. During this period, the rapid increase and widespread distribution of data reached an unprecedented level, and artificial intelligence (AI) algorithms became the key driving force to unlock the potential value of this data, leading various industries towards digitization and intelligence. However, in the process of data analysis and processing, security issues are like hidden currents that continue to invade the confidentiality and personal privacy of information. In recent years, differential privacy (DP) models have gained widespread attention and praise in both academia and industry due to their stable theoretical framework and excellent privacy protection effectiveness. Faced with the many challenges encountered in the field of data anonymization, this article proposes a unique solution: a data anonymization algorithm that combines differential privacy and deep learning (DL). This algorithm aims to utilize the powerful capabilities of DL to enhance the efficiency and accuracy of data anonymization, while relying on differential privacy technology to build a solid privacy protection barrier for data. The experimental results show that the algorithm can effectively protect personal privacy while ensuring the availability and accuracy of data.