Abstract: With the improvement of the quality of clinical diagnosis and treatment, the traditional scheduled "ward round" mode cannot meet the demands for real-time monitoring of acute and critically ill patients. This paper introduces the Storm, a real-time data stream processing technology and its application in the real time disease early warning system. By collecting the clinical data flow and calculating the MEWS scores in real time, the system can identify the potential deterioration of the disease, and promptly notify the medical staff. Score calculation results can be stored for further analysis and presentation as well.