<p>Hydraulic jumps play a crucial role in hydraulic engineering and river management. They influence sediment transport, erosion, and water flow dynamics in various water systems. The length of these jumps is a key factor in structural design. This study investigates the length of hydraulic jumps on smooth and rough sloping beds using laboratory experiments with various weir types and Froude numbers. Roughness was introduced using randomly placed Billy balls (1.59&#xa0;cm diameter). The pre- and post-jump depths were measured, and the hydraulic jump length was determined. Experimental data were used to develop new empirical relationships incorporating Manning’s roughness coefficient (n). Artificial intelligence has become increasingly popular for modeling hydraulic phenomena, including jump characteristics. In this study, computational models, including a Cascade Forward Artificial Neural Network (CFANN), Gene Expression Programming (GEP), and linear regression (LR), were trained and validated against the experimental data to predict hydraulic jump lengths on sloping smooth and rough beds. Sensitivity analysis revealed that the upstream Froude number and Manning coefficient are vital parameters for determining the jump length, with the latter being particularly significant. Gene expression programming outperformed artificial neural networks and traditional statistical methods in modeling jump length. A Cascade Forward Artificial Neural Network (CFANN) and regression model were found to be reliable tools for estimating hydraulic jump lengths. The proposed CFANN model demonstrated superior accuracy, achieving an R2 of 0.929 and other favorable performance metrics. The study developed new empirical relationships based on Manning’s coefficient using experimental data, which aligned well with previous observations on hydraulic jump length. The findings were consistent with earlier research, further validating the study conclusions and contributing to a broader understanding of hydraulic jump dynamics.</p>