Drilling state monitoring while drilling is intended to describe drilling modes, judge the operating performance timely, and analyze the causes of non-optimal modes so as to guide the driller to adjust parameters in time to ensure efficient, safe and stable drilling. Therefore, it is of theoretical and practical value to study drilling mode identification based on state monitoring while drilling. In this paper, the support vector machine is used to identify the drilling mode based on monitoring data. According to the identification results, drilling efficiency is evaluated, and factors affecting the efficiency are discussed to improve drilling performance. Finally, simulation experiments are carried out with the practical drilling data to verify the feasibility and effectiveness of the proposed method.