<p>To predict the behaviour of rocks in their natural conditions, it is essential to perform laboratory tests on the rock samples and obtain the various parameters such as uniaxial compressive strength, dry density, porosity, ultrasonic pulse velocity and slake durability index (SDI). The SDI test is an important test to determine the long-term effects of the environment on a rock structure, but it is relatively more time-consuming. This study highlights three types of rocks, namely sandstone, shale, and bituminous coal and uses the properties that are routinely performed in most geotechnical engineering projects to estimate the value of SDI indirectly to save time and effort. Linear regression, multiple regression and soft computing models were used to determine the best possible methods for estimating the SDI. The regression analyses were used to find the most reliable parameter(s) and how the availability of different number of parameters affects the predictability of SDI. Furthermore, the analysis was reinforced with the soft computing techniques, namely, Support Vector Regression Machine (SVR), Random Forest (RF), Extreme Learning Machine (ELM), and Extreme Gradient Boost (XGB). Separate prediction models were developed in MATLAB (v2021) using the soft computing techniques for both cycles of the slake durability test, and the performance of each model was assessed by the performance indices, namely, R², mean absolute percentage error, mean square error, root mean square error, mean square error, weighted mean absolute percentage error, variance account for, expanded uncertainty and t-statistic. A detailed comparison of the results was performed using rank analysis and heat maps.</p>