With the growing demand for indoor positioning, navigation algorithms are gaining increased research significance. Positioning based on inertial devices has become an interesting research hotspot due to its independence, short-term high precision, strong real-time performance and good compatibility with smartphones. However, the problem of long-term cumulative errors in navigation algorithms that rely on inertial sensors still exists, especially the problem of heading angle dispersion. To address this issue, we introduce a pedestrian heading angle updating algorithm based on inertial sensors and magnetometers. The algorithm determines the heading angle by considering magnetic field strength and geomagnetic inclination and integrates it with the gyroscope-calculated heading angle. The fusion weight will be adjusted based on the strength of abnormal magnetic field interference in the surroundings. We conducted the experiment in the corridor of the Navigation Laboratory. Results demonstrate that the proposed algorithm has better accuracy and robustness than other traditional positioning algorithms. In addition, the proposed system shows good versatility and practicality in real-time positioning with low cost and computational complexity.