Composition of virtual-real worlds has been widely used in many domains.Especially to the industry in recent years,the typical human and mechanical interaction (HMI) has gradually been unable to meet the needs of current technological development. This thesis proposes an intelligence control system including robot control theory,deep learning for image recognition and gesture interaction.We call this system as Alpha Robot Control System (ARCS) .There are three main contributions. (1)The intelligent control model is proposed by combining the forward kinetics (FK) and the inverse kinetics (IK). They charge for the manual control and the automatic respectively. (2) The deep learning algorithm is proposed for image recognition. It is used to identify the goods and sort them in ARCS. (3) The hand gesture control is implemented through the smart hardware called leap motion. In the experiment, the results demonstrated that the proposed ARCS establishes a virtual platform for the next generation of human-machine interaction in digital twins for robot control.