Relative soil moisture (RSM) is conventionally observed only three times per month. The low frequency cannot meet the data requirements for debris flow forecast. To supplement RSM data series, we fit RSM and antecedent precipitation index (API) by three methods from June 28th to Sep. 28th 2008 of Zhejiang Province. Direct fitting RSM and API (DFRA) is the best method of the three for the least relative error and absolute error. Segment statistics includes two different segment criterions. One is based on API, and the other is based on RSM .The third method is the least square method. The results of segment statistics based on API (SSBA) and the least square of API and RSM(LSAR) follow the DFRA. Segment statistics based on RSM (SSBM) has poor result for its major error. The average correlation coefficient of DFRA is 0.66. The average absolute error is 4.9 and the average relative error is 5.63%. The DFRA method can be used to supplement RSM data series even in the adjacent areas where there are no RSM data.