Recent natural disasters like flooding and slope collapse have shown the need for natural risk management system, as they endanger directly public health and cause severe damages on the national economy. In order to improve the efficiency of risk management systems, this management system based on AANN(Auto-Associative Neural Network)is proposed in this paper. AANN can be effectively used for identification of abnormal data and data compression. The proposed AANN-based risk management system collects and stores measurement data from sensors and transmits them to remote server for web-monitoring. Generally, it is desirable to transmit the compressed data instead of raw data in normal state. However, if dangerous situation happens, rapid tramission of measurement data should be required. These requirements are easily satisfied by using AANN. In order to verify the feasibilities of the proposed system, The AANN-based risk management system is applied to slope collapse monitoring system.