Residual water content (RWC) is an important parameter for the study of seepage theory and strength theory of unsaturated soils. However

it is very difficult to determine experimentally from soil-water characteristic curve (SWCC) because the residual state can not be reached normally. In this case

the RWC is often estimated using the empirical methods including the model fitting

so the rationality of these methods is worthy to be studied. After Wuhan clay samples with different initial void ratios were prepared

SWCCs were measured by using the pressure plate apparatus and the RWCs were calculated by using the model fitting method. Water evaporation test was carried out in the natural state. Subsequently

the residual time was defined according to the rate of water loss and was used to determine the RWC. The micro pore characteristics were studied using the nuclear magnetic resonance (NMR) technique

which was used to explain the microscopic mechanism of the RWC. The results show that the model fitting method can be used to estimate the RWC

but its accuracy is directly related to the choice of model and the restriction of preliminary range in RWC. Water evaporation test is a direct and feasible method to determine the RWC. The micro pores of Wuhan clay are distributed in triple peak pattern

and RWC is closely related to the water content in those pores before the first peak. According to the T2 spectrum area with the relaxation time of less than 0.267 38 ms

the RWC can be predicted accurately

however

this method needs further demonstration and improvement for other type of soil.