Objective:To develop early warning systems for unplanned extubation in critical patients to give the early warning and interference in order to reduce the risks of unexpected drawn tubes and guarantee the security of the patients.Methods:Took the model of risk warning system for unplanned extubation in critical patients as the core, obeyed the guide of the nursing intervention program, based on the information system of patients in ICU and relied on computer science technology to design the architecture for the interface layout, the partitions, the modular structure, the content and the function of the system. This warning system was designed from multiple angels and directions. This research retrospectively analyzed the 18 cases for unplanned extubation from January to December in 2016 using this system to verify its warning effectiveness.Results:The prototype of this warning system included login module, evaluation module, decision module, warning module, directive module and the module for canal fixed scheme. The system automatically determined the level of risk for critical patients to implement the layered pre-warning and screen the high-risk patients. Finally, individual nursing interfering method could be supplied. The detection rate of this warning system was 88.89% after retrospective analysis.Conclusion:The warning system for unplanned extubation patients is an automated, intellectualized and informationalized platform. It can effectively warn the high risk of the extubation patients and evade the risk of canal nursing.