Abstract: To address poor accuracy in existing track matching methods caused by inertial guidance data error accumulation, this study proposes an indoor pedestrian track matching algorithm using smartphone inertial sensors. The method combines inertial data characteristics with indoor environmental features to establish a targeted constraint criterion for track matching. An adaptively deformed oriented bounding box (OBB) corrects initial trajectories and independently adjusts subsequent paths. The algorithm analyzes diverse handheld postures to enhance adaptability. Experimental comparisons demonstrate that the proposed method performs robustly across trajectory types, increasing accuracy growth rates by up to 46.16% over existing approaches while meeting time-efficiency requirements for matching tasks.