Abstract: Abstract: Soil water diffusivity is one of the most important hydraulic properties during water transport in soil. It is very necessary to accurately estimate the key parameter in agricultural production. In this study, an analytical-correction approach was constructed to improve the accuracy of the soil water diffusivity using the horizontal imbibition, particularly considering the boundary effect. Assuming that the one-dimensional horizontal imbibition (Richards's equation) was similar to the analytical solution of the linearized equation under the first types of boundary conditions, the approximate analytical solution was deduced using the constant variation. Hence, two parts were divided for the time-dependent change of the water content at the measuring point in the finite-length soil column, i.e., with/without the boundary effect. Firstly, two parameters were determined to approximate the analytical solution without the boundary. Secondly, the experimental data were corrected by the analytical solution considering the boundary effect. Finally, the soil water diffusivity was determined to analyze the corrected data using the horizontal imbibition. As such, an analytical-correction approach was established to estimate the soil water diffusivity, further to verify the numerical and physical laboratory experiments. A Hydrus-1D software was used to simulate the migration process of water in six soil textures, including the loam sand, sandy loam, sandy clay loam, loam, silt clay, and silt. An indoor horizontal imbibition test was also carried out to verify the analytical solutions and analytical-correction model. The types of test soil were taken from the topsoil of ionic rare earth mines in Xinfeng County, Jiangxi Province, Xunwu County, Jiangxi Province, and Pingnan County, Fujian Province. The volumetric water content with time and coordinates were real-time monitored at the two measuring points (p1 and p2,) using the FDS-100 water sensor. The results show that the analytical-correction approach can be widely expected to accurately correct the data caused by the boundary effect, where the closer to the water inlet was, the larger the corrected water content range was. Both the laboratory test and Hydrus-1D simulation demonstrated that the determination coefficients of the soil water diffusivity were all above 0.900 using the new analytical-correction, which was much higher than that of the horizontal imbibition only (below 0.600), although the time-varying data of water content was susceptible to the boundary effects. Furthermore, the coefficients of determination were all greater than 0.950 for the test at the different measuring points in the indoor test. Consequently, the new analysis-correction can be expected to accurately estimate the soil water diffusivity, particularly combining with the horizontal imbibition.