Abstract: OBJECTIVE By application of ATP bioluminescence method in detection of the quality of surface cleaning and disinfection in ICU, to investigate the effect of cluster intervention, so as to prevent the spread of multidrug resistant bacteria. METHODS From Jun. 2013 to Jul. 2016, the cleaning and disinfection effects of the surfaces which ICU medical staff in a tertiary hospital touched with high frequency by hands(various types of instrument panel and surface, all kinds of countertops, door handles, computer mouse and keyboard, etc.) were sampled and monitored, and were detected by ATP biological fluorescence from Jan. 2015. Cluster intervention measures were implemented in ICU. The cleaning and disinfection effects before and after the intervention were compared, SPSS16.0 software was used for statistical analysis, and the rates were compared by χ²-test. RESULTS After the use of cluster intervention measures, the qualified rate of surfaces in ICU by ATP biofluorescence was increased from 20.8% to 44.0%, and the difference was significant (P<0.05).A total of 147 samples were sampled. Using ATP scores <100 RLU/100 cm² as cleanliness criteria, and classifying by environmental object surface, the qualified rates of instrument panel and surface, countertop, door handles, computer mouse and keyboard, telephone, and other surfaces were 33.3%, 52.4%, 33.3%, 15.5%, 0%, and 38.5%, respectively. The top three sampling sites with the worst surface cleanliness of the four types of objects were 377~9740RLU/100 cm². CONCLUSION ATP bioluminescence monitoring data show that, the effects of application of cluster intervention in the cleaning and disinfection of ICU environmental object surface are significant, but the overall cleaning effect is still not ideal, and it is the important hidden danger for hospital infection management and multidrug resistance prevention and control. How to strengthen routine cleaning and disinfection work to improve the overall surface cleanliness of environmental objects still needs to explore.