The raster map generated by the particle filter-based SLAM algorithm has the problem of low storage efficiency. To address this issue, a SLAM algorithm based on Huffman run-length coding is proposed. The coding redundancy and space redundancy problems of the original raster map are solved. Based on the particle filter-based SLAM algorithm, the algorithm uses Huffman run-length coding map representation. According to different application scenarios, two storage methods of Huffman run-length coding maps are proposed. When the scale of the grid map is small, the fixed-length coding is used. When the scale of the raster map is relatively large, such as a large shopping mall environment, the variable-length coding is utilized, which further expands the application range of the map representation. Simulation and real scene experiments show that, under certain conditions, the SLAM algorithm based on Huffman run-length coding can reduce memory consumption by up to 94. 8% , which proves the feasibility and effectiveness of the algorithm.