Precise estimation of unsaturated soil hydraulic conductivity is essential for designing critical infrastructure, including foundations, retaining walls, and underground structures. Such estimations enable engineers to evaluate soil stability, forecast seepage behaviour, and develop effective drainage solutions. Traditional approaches for estimating unsaturated soil conductivity often depend on intricate theoretical models or empirical equations, which may be limited in accuracy and practical applicability. This study employs random forest regression (RFR) and artificial neural networks (ANNs) to model and predict unsaturated hydraulic conductivity, utilizing key input parameters such as soil suction and particle size distribution. Results reveal that the RFR model outperforms other methods, delivering superior predictive accuracy with a minimized mean absolute error (MAE). Comprehensive performance evaluation is conducted using metrics including root-mean-squared error (RMSE) and coefficient of determination (r2), highlighting the efficacy of the proposed predictive models.