During the process of indoor positioning, the map matching algorithm based on particle filtering can effectively fuse the indoor positioning results with the map data. Currently, this method mainly faces two problems: the calculation complexity of the wall-crossing detection algorithm during map matching and issue of the positioning results being in unreachable areas. To address the issue of high calculation complexity in the wall-crossing detection algorithm, we propose a wall-crossing detection algorithm based on a map information matrix on a grid map, which reduces the calculation complexity while ensuring the correct detection of wall-crossing particles. To address the issue of positioning results being in unreachable areas, we propose a map matching algorithm based on particle filtering with multiple weight updates to correct positioning results in unreachable areas. Theoretical analysis and experimental results demonstrate the rationality and effectiveness of the improved algorithm.