The invention belongs to the technical field of big data systems, and particularly relates to a method and device for monitoring a state of a big data system, and the method comprises the following steps: 1, enabling a system to access big data, and collecting data information; 2, splitting the collected data information according to time points; 3, performing time point comparison according to the split data; 4, performing data correction on the compared information; 5, combining the corrected data, and then performing data delay elimination; and step 6, outputting the data of which the delayis eliminated to a system state monitoring system. Data splitting is carried out on collected data system state information; the split data are compared and corrected, and then the corrected data aresubstituted into the current time point to be synchronized and then output to the system state monitoring system, so that the delay effect of system state monitoring under big data is effectively solved, and loss caused by data updating errors is prevented.