The spatial distribution of soil organic matter (SOM) has intimate connection with topography, but the synergy effect of topography for SOM spatial redistribution was frequently ignored in the study of SOM by traditional geostatistics. In order to investigate the influence of different topography on the spatial distribution of SOM, the method of SOM spatial interpolation based on topographic unit was proposed. Firstly, the terrain factors such as slope, aspect, topographic wetness index and topographic position index were calculated from the digital elevation data (30m×30m). Secondly, the topographic units were precisely divided into 13 different classical types more subtly by integrating the terrain factors, which effectively avoided the traditional terrain classification method that only based on elevation data. The regions were divided, which was based on terrain classification rules formed by the distribution of terrain factors in different landforms. Thirdly, soil samples were collected in different topographic types, and the distribution of SOM for each sample set in different topographic units was generated by ordinary Kriging. Then, the corresponding results of interpolation for each sample set were segmented based on topographic unit region, and combining the result in each region, the spatial distribution of SOM based on topographic unit was obtained. Finally, verification and comparison with the accuracy of each SOM distributions were performed, which were obtained by using topography based geostatistics and traditional global geostatistics, respectively. The results of experiment for researching the SOM distribution in Zhongxiang City showed that higher accurate and efficacious spatial layout of SOM can be acquired by geostatistics based on topographic units, especially in the gently topography region, the precision was increased by 16.39%. The SOM prediction accuracy in the rest of the terrain unit was similar to the global prediction accuracy with Pearson correlation coefficient of 0.75. Using geomorphic zoning to obtain higher accuracy for spatial distribution of SOM, and the influence of topography was considered in the study of SOM geostatistics. 