High-precision positioning of a shearer is an important research direction of automation and intelligence in coal mining. The inertial navigation system and odometer are the main sensors for positioning of the longwall comprehensive mechanized shearer, which can effectively suppress the divergence of the inertial navigation system and has better autonomous navigation ability through information fusion. However, it still cannot meet the requirements of long-term and high-precision navigation. This paper analyzes the existing problems of auxiliary sensors in the mining process of shearer and puts forward an improved factor graph optimization method based on UWB measurement at both ends of the shearer working face. The constraint equation and graph optimization model of the inertial navigation system/ odometer/UWB are derived and constructed by position measurement information of UWB. At the same time, through the pre-integration of inertial information, the number of nodes to be optimized is reduced, thereby reducing the calculation amount of the proposed algorithm. On this basis, the factor nodes of the odometer scale factor error and installation error are added for joint estimation and optimization. Finally, simulation and actual experiments are carried out. The results show that compared with the traditional Kalman filter, the proposed algorithm effectively improves the positioning accuracy of the shearer.