In order to solve the problems of inflexible robot grasping action and low operation precision caused by the single built-in sensor,a robot grasping system based on multi-sensor was proposed.The extended Kalman filter algorithm was used to fuse the position,speed and angle information measured by the robots built-in sensors to realize the robots own pose detection.Under this premise,an external sensor was used to complete the automatic identification and positioning of the target.In order to enhance the reliability of information and the distinguishing ability of the system,and perform kinematics modeling of the gripping action of the robot according to the output signal,the hardware and software environment of the gripping system and the transmission circuit were set to complete the whole system design process.The experimental results prove that the robot system designed by the proposed method can achieve precise alignment to the gripping target,the motion characteristics have been significantly improved,the operation error is effectively reduced.