With the overall improvement of the digital and intelligent level in the nuclear power industry, the failure of the Digital Control System (DCS) has become one of the major factors affecting the stable operation of nuclear power units. DCS safety problems can easily lead to the unit outage or shutdown, power reduction operation, system or equipment degradation events and other serious consequences, which will significantly threaten the safety and reliability of the unit operation. This paper studies the real-time safety problem of DCS based on the monitoring data of DCS itself, focusing on DCS current operating risk state identification, operating risk state transition model construction, residual operating risk prediction, to evaluate DCS operating risk state and predict the risk evolution trend. Furthermore, this study focus on the design and development of the software tool for risk state identification and prediction of DCS, so as to implement the methods and algorithms presented herein.