Abstract In order to build a navigable high-precision gravity reference map efficiently, it is indispensable to improve the precision of the interpolation algorithm, and take its optimization into account. In view of the difficulty of optimizing an efficient interpolation algorithm for gravity reference maps in the various interpolation algorithms, a interpolation optimization algorithm for gravity reference maps in local anomaly fields is put forward. This algorithm realizes preliminary optimizing of the different algorithms by studying long-term data interpolation, determining the weight by means of an analytic hierarchy process, and defining quantitative coefficients of the optimization algorithm (named optimizing degrees). The experimental results show that the optimization algorithm is operational and that the time to interpolate a gravity reference map is reduced in special conditions.