Abstract: The soil-water characteristic curve (SWCC) is an important curve describing the relationship between the suction of unsaturated soils and the saturation or water content, and is an important basis for analyzing the strength, deformation and seepage of unsaturated soils. It is very time-consuming to measure soil suction directly or indirectly indoors. In order to quickly and accurately obtain the SWCC of unsaturated soils, an improved method for predicting SWCC based on the pore size distribution (PSD) of soils is proposed. This method uses the mercury intrusion porosimetry (MIP) to measure the PSD of soils, and the filter paper method is used to measure a suction value of soil samples and its corresponding saturation. Then according to the test results at this point, the pore volume of soils measured by the MIP tests is corrected, and the pore volume after correction is used to calculate the saturation of soils under different suction conditions. This method can overcome the problem of small pore volume measured by the MIP tests. The comparative analysis of the predicted and measured SWCCs before and after the correction of 9 groups of soil samples shows that the proposed method can predict the SWCC of unsaturated soils more accurately. On this basis, the fitting parameters of SWCC and their probability statistical characteristics of multiple groups of soils can be obtained conveniently and quickly.