This paper describes a framework for highly distributed real-time monitoring approach to Database Security using Intelligent Multi-Agents. The Statistical Anomaly Prevention system employs Back-Propagation Neural Network forecasting model, which predicts unauthorized invasions of user based on previous observations and takes further action before intrusion occurs. Our Back-Propagation Neural Network model makes periodic shortterm forecasts, since long-term forecasts cannot accurately predict an intrusion. We use a multivariate time series technique to forecast the hacker's behavior effectively. Our Back-Propagation Neural Network model results have been compared with traditional statistical forecasting models, and a better prediction accuracy has been observed. In order to reduce single point of failures in centralized security system, a dynamic distributed system has been designed in which the security management task is distributed across the network using Intelligent Multi-Agents.