Hospital network security dynamic control technology is of great significance to ensure the security and stability of hospital networks. The traditional network anomaly monitoring and network security dynamic control cannot solve the problem of large area network intrusion. Therefore, in order to solve these problems, the study constructs a hospital security dynamic control model based on Self-Organizing Incremental Neural Network (SOINN) algorithm combined with Advanced Digital Network Data Design (ADNDD), a digital information processing network technology. ) for hospital safety dynamic control model. Firstly, the algorithm is optimized, secondly, the SOINN and ADNDD are fused to construct the network security dynamic control model, and finally, the performance of the model is verified using the dataset. The results show that after training in the dataset, the outlier recognition rate of the model in the datasets of surge attack, deviation attack and geometric attack is 92.13%, 90.04% and 89.07%, respectively. This indicates that the model algorithm can meet the requirements of network security in hospital network anomaly detection and dynamic defense control after the application of the dataset. The aim is to improve the security and stability of hospital networks.