Abstract: Aiming at the problem of the mutual contradiction between positioning accuracy and calculation amount in the high-precision underwater acoustic positioning system, an adaptive stratification algorithm based on minimum variance for acoustic ray tracking is proposed in this paper. According to the characteristics of regular variation of sound velocity in a limited range and by the global search of the sound velocity profile, the total variance from the dispersed sound velocity points of each segment to the fitted line is used to represent the difference between the segment to be segmented and the original sound velocity profile. The best segmentation point is found by the maximum difference, and the characteristics of the original sound velocity profile are retained to realize the simplification of the sound velocity profile. The experimental results show that the proposed method can increase the computational efficiency while ensuring the positioning accuracy, and has good engineering application value.