Abstract: Objective To analyze the application of automatic early warning system for commun icable diseases for improving alert efficiency.Methods With dynamic percentile and one case early warning method, the early warning of communicable disase system was estab lished.Data from the system were analyzed with epideiologic method.Results In 2009, the system issued 1964 times of a lertinformation with a positive ratio of 0.20 percent.The regional distribution of the information was different significantly, showing a positive correlation between the number of disease case and the time of early warning(χ²=56.15, P < 0.01; r=0.84, P < 0.01).The distribution of the information for different types of the diseases was also different significantly, showing a positive correlation between the prevalent strength of the disease and the amount of information in a specific area(χ²=195.67, P < 0.01; r=0.64, P < 0.01).Conclusion Positive predictive valueislow, but the sensitivity is high for the early warning information system.