With the characteristics of unmanned, long-term operation, a large number of devices and complex interactions, itis crucial to assess the health of ground-based monitoring system for satellite signal monitoring system. By reasonably designing andprocessing a large amount of operational data from the monitoring system, a regression neural network to learn is designed andtrained and the inherent laws of the monitoring system’s operation data is explored. The enables intelligent monitoring and faultwarning throughout the entire lifecycle of the monitoring system, shifting the maintenance mode from post-event to pre-event,reducing the occurrence of monitoring system failures, and improving the stability and reliability of the system.