Abstract: Objective To Develop the ICU nurse's KAP scale in the general hospital and test its reliability and validity, in order to assess the subsyndromal delirium in ICU. Methods The subsyndromal delirium scale framework and items was initially built through literature reviewed and integrated，combined with expert opinion and knowledge theory. The reliability and representativeness of expert consultation are tested with expert authority coefficient, positive coefficient, coordination coefficient and coefficient of variation. After the formation of the questionnaire, we investigated 510 nursing staff working in ICU at 7 tertiary A hospitals and 2 tertiary B hospitals in Zhejiang province from July to August 2017, with the above questionnaire, filter items by project differentiation analysis, correlation analysis, exploratory factor analysis prevention; Cronbach's a coefficient, retest reliability, content validity and structural validity analysis were used to determine the feasibility and validity of the questionnaire. Results The participation in expert consultations is 100% and the authoritative coefficient is 0.87, the consistency of the experts in the two rounds of correspondence is 0.75 and 0.89. The ICU nurse's KAP scale in the general hospital contains 3 dimensions of knowledge, attitude and care behavior, including 32 items, using Likert 5-grade scoring method. The total Cronbanch's α coefficient of the questionnaire is 0.964, the Cronbanch's α coefficient for each dimension is 0.900-0.976, the test reliability is 0.985, the retest reliability for each dimension is 0.836-0.952, the content validity indicator S-CVI is 0.97, I-CVI is 0.97-1.00; the cumulative contribution rate for each dimension is 69.38%.The confirmatory factor analysis shows that the questionnaire has a good structural validity. Conclusion This scale has good reliability and validity, and can be used to investigate the status of subsyndromal delirium of nurses in ICU.