The objective of this research was to investigate the relationship between soil moisture characteristic curve (SMCC) and confined compression curve (CCC), and the ability to estimate the SMCC from the CCC. Five provinces of Iran have been chosen as sampling sites and soil samples (150) were collected from these areas, and some properties of the soil samples were determined. The Gardner model has been fitted to the measured SMCC and CCC. The Gardner model's stress-void ratio coefficients, as well as the CCC properties, were used to estimate the soil moisture at five levels (five classes of variables) through the Gardner model using artificial neural networks (ANNs). A more accurate estimation of the water content was obtained by combining the basic soil properties, and three key properties of soil compression. In addition, the integral root mean square error (IRMSE) in the training and testing steps was reduced from 0.107 and 0.111 to 0.095 and 0.096, respectively. In conclusion, the use of CCC data to estimate the SMCC at all levels of the input variables indicated very favorable results with an increase in the accuracy of the water content estimation between 4% and 16% in both training and testing steps.
