<p>The accurate prediction of bedload transport is crucial in designing and managing alluvial channels. However, many empirical models have limited or no evidence that has been collected from controlled studies, especially for low gradient, shallow sand bed flows. Therefore, this study has examined bedload transport behavior through a series of controlled laboratory experiments which utilized fine non-cohesive sand within a 20-m-wide rectangular tilting flume. A full-factorial experimental design comprising 18 runs was implemented by varying channel slope, discharge, and bed thickness. Through systematic measurement of both flow characteristics and transported sediment, this study quantified both hydraulic responses and bedload transport rates at subcritical turbulent flow conditions. Results showed a highly nonlinear relationship between bedload transport and channel slope and discharge while bed thickness had a secondary effect on sediment mobility conditions. Overall, the classical Schoklitch equation performed well relative to experimental data resulting in a mean error of approximately 12%. As a result, deviations increased with higher flow conditions. Response surface methodology (RSM) demonstrated excellent predictive ability (<i>R</i>² = 99.84%) and identified channel slope as the primary factor controlling bedload transport rates. Additionally, artificial neural network (ANN) successfully captured the complex nonlinear inter-relationships between these parameters and resulted in superior predictive accuracy than empirical modeling. The study provides a comprehensive experimental dataset and a unified framework integrating empirical, statistical, and machine learning approaches for bedload prediction. These findings contribute to improved understanding of sediment transport mechanisms in shallow open-channel flows and offer reliable tools for sediment management and hydraulic engineering design.</p>