Objective To construct the algorithm of a nosocomial infection early warning system (NIEWS) with application of Large Data Screening based on the real-time nosocomial infection surveillance system (RT-NISS). Methods Large Data Screening and Mathematical Model were adopted to design a series of algorithms. Through process data acquisition and analysis, the influences of various influencing factors were compared, and then given different weights. The influencing factors for individual patient were also given weights. Then the warning interval was set to implement the early warning of nosocomial infection. Results Through the data mining of the Infection Information Database from the RT-NISS, a series of algorithms were successfully designed, and the risk factors of nosocomial infection were screened out. Then the weights of these risk factors were verified and adjusted in combination with the cases of nosocomial infection from hospitals. The early warning interval was set and implemented for nosocomial infection. ConclusionThe preliminary framework of the NIEWS based on RT-NISS system is designed and established. In the further study, we will verify and compare in combination with nosocomial infection data in order to improve the accuracy and precision of prediction, and implement the early warning before the infection, induce clinical concern and take preventive measures to reduce nosocomial infection rate.