The invention relates to a method for identifying a working state of an action track mode in the field of big data processing, which comprises the following steps of: acquiring employee track data andcleaning the track data; extracting the node number, the track node number, the node access frequency and the track total length with long residence time according to the track data; training historical track data through a logistic algorithm, an svm algorithm and a decision tree algorithm; and repeatedly adjusting and optimizing the model, and identifying the new track data. The attendance condition of the staff can be well judged, and the working property can be judged according to the behavior track information of the staff.