Abstract: Aiming at the difficulties of mobile robots in the underground tunnel of a coal mine, such as difficult position sensing and low positioning accuracy due to the influence of the environment, a precise positioning method for mobile robots using laser targets as the positioning reference is proposed. According to the characteristics of the roadway environment and the laser tracking detection principle, establish the distributed deployment strategy of the laser target network, construct the body positioning detection platform based on the laser target network and laser tracker; based on the known spatial information of the roadway and the laser target position information, construct the roadway coordinate system and the position information database and build it into the laser target, and utilize the positioning detection platform to decipher the actual position information of the laser target under the laser transmitter coordinate system, and establish the laser target positioning accuracy. The actual position information of the laser target in the coordinate system of the laser transmitter is solved using the positioning detection platform, and the correlation relationship between the coordinate system of the laser target in the roadway and the coordinate system of the fuselage is established; the coordinate information of the four characteristic points of the laser tracker stored in the database is used as the input of the positioning model of the laser transmitter, and the laser target, and the position information of the fuselage is solved by the EPnP (Efficient Perspective-n-Point) algorithm, and the roadway of the laser target is used to solve the position information of the fuselage. The body position information is solved by EPnP (Efficient Perspective-n-Point) algorithm, and the minimum residual sum of squares of the laser target lane coordinates and the measured coordinates is used to characterize the body position error, construct the body position error model, and minimize the body position error by Levenberg-Marquardt algorithm, and obtain the accurate position information of the mobile robot in the lane coordinate system through the spatial coordinate conversion. After experimental verification, the maximum error of mobile robot distance measurement along the direction of the roadway, the width direction, and the height direction is less than 11.33, 8.12, 8.57 mm; the maximum errors of the actual and theoretical positions of the dynamic trajectory measurement in the roadway direction, width direction and height direction are 2.74%, 3.14% and 2.36 %, and the stable measurement of body positioning is guaranteed in the complex roadway with turns. The stable measurement of body positioning meets the requirements of positioning accuracy of mobile robots in underground coal mines.