AbstractThe permeability coefficient of unsaturated soil and the soil–water characteristic curve (SWCC) plays a significant role in the theory of unsaturated soil mechanics. The pressure plate can quickly obtain the SWCC. Still, the unsaturated permeability coefficient often needs more expensive test equipment or more complex methods to be measured separately, which is time-consuming, laborious, and costly. In this paper, the pressure plate, a standard instrument for measuring the SWCC, is used as the primary measuring device. By using different ceramic plates with various air intake values and changing the sample size, the drainage equilibrium time under each level of matric suction is controlled by the permeability characteristics of the sample. Meanwhile, the reading device of the pressure plate meter is improved. Finally, a new method for synchronous measurement of unsaturated soil permeability coefficient and SWCC based on the axial translation technique is put forward. Taking Hunan clay as the research object, the SWCC and unsaturated permeability coefficient of clay samples with three dry densities were measured. The measured SWCC data points were fitted and analyzed by Fredlund and Xing’s (F-X) model and Van Genuchten’s (VG) model. The Childs and Collis-George’s (CCG) model and Tao Gaoliang and Kong Lingwei’s (TK) model predicted the unsaturated permeability coefficient, and the anticipated results were compared with those obtained by the new method. The results show that the new approach can simultaneously measure the SWCC and unsaturated permeability coefficient of Hunan clay. In the range of matric suction (<200 kPa) studied in this paper, the larger the dry density of samples, the smaller the change range of water content with matric suction. Under the same matric suction, the greater the dry density of the soil, the longer the drainage equilibrium time, and the smaller the unsaturated permeability coefficient. The F-X and VG models have a good fitting effect on the measured data for SWCC. The predicted value of the TK model is close to the measured result of the new method, and the predicted error of the CCG model is large.