In this paper, an intelligent handling robot system integrating laser SLAM (Simulated Localization and Mapping) navigation and color recognition functions is proposed, aiming at solving the requirements of material sorting and transportation according to color classification in complex dynamic environment. The system adopts multi-sensor fusion architecture, realizes high-precision environment modeling and target detection through 16-line lidar and RGB-D camera, and combines the improved YOLOv5s color recognition algorithm and dynamic path planning strategy, which significantly improves the sorting accuracy and working efficiency. Experiments show that the positioning error of the system is less than 12 mm, the accuracy of color recognition is 98.7%, and the single task cycle is shortened to 25 s, which is 45% higher than the traditional scheme.