The Automated Guided Vehicles (AGVs) on the production lines mainly utilize positioning and navigation technology based on Light Detection and Ranging (LiDAR). There are problems with inaccurate positioning data due to data drift, environmental interference, and error accumulation. Based on Ultra-Wideband (UWB) sensors and data processing technology, the paper proposes a UWB sensor-assisted indoor localization method for AGVs in intelligent manufacturing factories. Three-dimensional spatial localization of the indoor production line is achieved by LiDAR and UWB sensors, which are used to acquire AGV coordinate information. During the operation of the AGV, the AGV communicates with the service end of the production line using the TCP communication protocol under the LAN to collect and store the coordinate data of the AGV in real time. Furthermore, this paper is based on statistical measurement principles and the “3σ principle”, preprocessing the data collected by the UWB sensor. The coordinate data collected by the two positioning methods are fitted and optimized using Kalman filtering. The experimental data are analyzed by trajectory analysis and calculating the average absolute error. The results show that the method in this paper controls the error range of AGV coordinates within ±0.2 meters, which meets the standard of indoor high-precision positioning. This paper combines LiDAR and UWB sensors with data processing techniques to provide a solution to the challenges associated with the high-precision indoor positioning of AGVs.