Lubricating oil monitoring is one of the important methods of condition monitoring and fault diagnosis of shield machine main bearings.However,the current diagnostic threshold values are mostly empirical data,and there is a lack of systematic analysis of a large amount of lubricating oil monitoring data.In order to establish a reasonable threshold for oil monitoring and fault diagnosis of shield machine main bearing,the oil monitoring scheme for main bearing of shield machine was designed,the sampling method,sampling period and sampling period under special circumstances for oil monitoring of shield machine main bearing were proposed,and a large amount of oil monitoring data were obtained.The benchmark value,warning value and critical value of the monitoring items were calculated by statistical method,which were consistent with the current monitoring threshold values established through empirical data,and the rationality of the current monitoring threshold values were proved theoretically.The regression analysis was performed on the lubricating oil monitoring data by linear regression method to explore the relationship between lubricating oil monitoring data and excavation mileage,so the lubricating oil monitoring threshold values of the main bearings in each stage of the life cycle were more reasonably determined,which complemented the threshold values calculated by the statistical methods.Through correlation analysis,the internal relationship of the data changes in the lubricating oil monitoring was identified,and it was pointed out that the water content in the lubricating oil was the primary control factor in the condition monitoring of shield machine main bearings.