Abstract: Qujiang, a typical hilly area in Southern China, was selected to analyze the spatial interpolation for soil organic matter. In the experiment, two spatial interpolation methods, ordinary Kriging (OK) and inverse distance weighted (IDW), were applied. The results showed that IDW was better than OK in the same data processing condition. Under one interpolation method, the comparison among 3 conditions including classifications according to terrain and physiognomy, farmland use type and non\|classification showed that the classification of farmland use type was superior to the classification of terrain and physiognomy, and both of them were much better than the non\|classification．

Key words: interpolation, precision evaluation, ordinary Kringing (OK), inverse distance weighted (IDW)