Abstract: Drilling and coring, as an effective method of acquiring deep lunar regolith, has been widely applied in extraterrestrial sampling missions. Different from drilling on Earth, unmanned lunar drilling & coring may meet several technical problems, such as time delays in remote control, limited sensor resources, lack of geological information on sampling site, complicated mechanical properties of lunar regolith and so on. To realize high efficient drilling process with high reliability and have adaptability on unknown drilling environment, sampling device should adjust drilling parameters online depending on the real-time drilling conditions by limited hardware resources on the probe. This paper proposed a control method of lunar drilling based on online identification of drilling ability. The intelligent drilling control method has been realized by using drilling ability index to describe the drilling difficulty level, adopting pattern recognition method to identify the drilling ability levels and matching the optimized drilling parameters online. In order to verify the proposed control method, the drilling experiment in a multi-layered simulation mixed with granular soil and hard rocks has been conducted. Experimental results showed that drilling load under this control method could be controlled effectively.