Abstract: In the context of intelligent fully mechanized caving and mining, it is important to consider not only the cutting by the shearer, but also the coal caving at the back of the support and top coal recovery. Compared with intelligent fully mechanized mining face, the characteristic features of intelligent fully mechanized caving face are having more equipment, more complicated working conditions, more possibilities of abnormalities in the production process and more manual remote interventions. In order to ensure the smooth operation of intelligent fully mechanized caving and mining, automatic operation and manual intervention analysis system for intelligent fully mechanized caving face is designed. The definitions of key indicators are defined, such as the automation rate of intelligent fully mechanized caving face and the operation rate of the caving face control system,the shifting rate of automatic following machine of the hydraulic support, shearer memory cutting rate and automatic roof coal caving rate. By collecting the operation information of the main equipment in fully mechanized caving face and the coal mining information, the operation status of each equipment is reviewed and the statistical analysis of automation rate of fully mechanized caving face is obtained. Based on the threshold knowledge information and equipment work-flow of the automatic operation status of fully mechanized caving face, threshold knowledge database of the automatic operation state change of intelligent fully mechanized caving face is established. The database is used to identify the status of production equipment before manual intervention and estimate whether it is suitable to stop automatic operation of production equipment. Based on the in-depth analysis of expert knowledge, a rule database is established. When implementing manual intervention, the database automatically estimates whether the conditions for removing automation are met based on the rules in rule database, and analyzes the reasons for manual intervention. The results show that the system is able to evaluate the automation rate of intelligent fully mechanized caving face, analyze the reasons for manual intervention, and provide a basis for optimizing the control logic of production system.