Abstract: Objective To understand the performance of automatic early warning information system (AEWIS), evaluate its operating results, and to provide scientific basis for further improvement of communicable disease early warning patterns. Methods Descriptive analysis method was used to describe the characteristics of warning information sent out by AEWIS from 2008 to 2013. Sensitivity, positive predictive value and lead time were taken as the main evaluation indicators for analyzing operating results of AEWIS. Results AEWIS issued 13 021 warning signals from 2008 to 2013 and 12 of the signals were confirmed to be outbreaks or epidemics through field investigations. Twenty-four kinds of infectious diseases were identified from the all early warning information issued. A total of 11 888(91.30%) diseases were hand, foot, and mouth disease, other infectious diarrhea, mumps, dysentery, hepatitis C, and measles. The peak number of early warning signals were 4 880 from May to July, accounting for 37.48% of the total signals. The median response time from early warning signal to submitting abnormal information card was 0.40 hour. The sensitivity of AEWIS was 85.71% and the positive predictive value was 0.09%. The lead time of the 8 of 12 outbreaks of infectious diseases detected by AEWIS was negative. Conclusion The sensitivity of the warning system is high, but the positive predictive value is low. It is recommended to adjust the threshold timely and reasonable, so that the early warning system can play a role in early warning outbreaks of infectious diseases at the grassroots level.