Abstract: The iterated closest contour point (ICCP) algorithm results in large matching errors and even false matching in areas with small geomagnetic variations. Given this, this study established an error model for the ICCP algorithm in order to determine the causes of the large matching errors in the areas. Based on this model, this study proposed an iterated search closest contour point (ISCCP) algorithm suitable for geomagnetic gradient matching for navigation. In the proposed algorithm, the closest point search method was improved based on the ICCP, and an iterative search was performed using the geomagnetic gradient information of three orthogonal directions. The simulation results show that the average point error of the ISCCP algorithm can be controlled within half a grid length, with a precision about 15 times higher than that of the ICCP algorithm. Therefore, the ISCCP algorithm proposed in this study can effectively eliminate the large matching errors in areas with small geomagnetic variations caused by the ICCP algorithm.