Abstract: With the development of artificial intelligence and multimedia technology, rotary tillers play an important role in agricultural production. It takes the rotary tiller motion control system as the core. It first introduces the overall struc-ture and control principle of the rotary tiller, then designs a multimedia speech recognition system based on MFCC feature extraction and VQ vector method, which can realize the control of the rotary tiller depth and motion. And then, it carries out relevant experimental research based on the system, which proves that the system can work normally, and the average recognition rate of voice control is above 94%. It can meet the voice control requirements of the rotary tiller, and it proves the reliability of the voice control system of the rotary tiller.