As geotechnical engineers going through their investigation, they need to know basic information about the unsaturated soil property function they are searching for. To find out more about the unsaturated soil property they are analyzing, they need to look through the soil-water characteristic curve. A soil-water characteristic curve is a tool that describes unsaturated soil properties, such as the air-entry value and residual suction. Predicting unsaturated soil property by direct measurement of soil-water char-acteristic curve (SWCC) using a pressure plate test consumes too much time. In such cases, predictive models can become a practical alternative to obtain an estimate of the SWCC. The main objective of this research is to develop a prediction model for sim-plifying SWCC parameters for the Fredlund & Xing (1994) model fitting parameters based on correlations with soil index properties. Twelve red clay soils were collected from different field sites of Assosa, which is located in the western part of Ethiopia. The soil samples were subjected to laboratory testing that included index testing and SWCC testing. SWCC data point was measured by pressure plate apparatus experimen-tally for the twelve red clay soil, and the measured data point was optimized to obtain the Fredlund & Xing (1994) model fitting parameters using the gradient-based algo-rithm of MATLAB ® R2020a function called “fmincon” routine. The linear regression was performed with the help of JMP ® Pro14 statistical software to correlate the opti-mized Fredlund & Xing (1994) model fitting parameters and index properties. To vali-date the proposed prediction model a cross-validation technique was used. To further validate the rationality of the prediction model a statistical analysis called monotonicity analysis, and robustness analysis was done. To develop the prediction model a data set of nine samples were used. While, the remaining three samples were used to cross-validate the proposed prediction model. This study demonstrates that the weighted plas-ticity index can be used to develop a robust prediction model for the Fredlund & Xing (1994) model fitting parameters with acceptable fit statistic and internal validity. In comparison with previous models; the proposed model was close to the actual SWCC curve and SWCC data points for each suction level.