Abstract: In order to solve the problems of a poor driving environment and low intelligence level of underground load-haul-dump （LHD） vehicle, a simulation study on autonomous positioning and navigation of underground intelligent LHD vehicle was achieved through kinematics analysis, 3D model establishment, global map construction, and global path planning. By analyzing the kinematics model of articulated vehicles, the correlation between vehicles’ physical parameters was established, and the conditions that the LHD vehicle simulation model needed to meet were clarified. On the Gazebo 3D simulation platform, a simulation model of the LHD vehicle body with sensors and a simulation model of the underground tunnel environment were constructed. The initial position of the LHD vehicle simulation model in the tunnel environment was set to facilitate subsequent autonomous positioning and navigation research. Based on sensor information, a global map was constructed. The AMCL algorithm was used to achieve autonomous positioning of the LHD vehicle simulation model, ensuring that the initial position of the LHD vehicle simulation model was consistent between the Gazebo 3D simulation platform and the Rviz 3D visualization platform. Two algorithms were used for global path planning of the LHD vehicle simulation model from the mining area to the unloading area in underground tunnels, respectively. The simulation results show that the A* algorithm performs better in this simulation study, providing a technical support for achieving intelligent underground vehicles.