Monitoring bedload transport is of interest for studying river morphology evolution and hydraulic structures stability (e.g., dams andbridges). Bedload self-generated noise measured by hydrophones has been experimentally correlated to bedload flux in several studies. However, the lack of theoretical background linking the recorded acoustic noise to bedload and river characteristics has prevented a good understanding of the experimental results. Here, we develop a model of the acoustic noise generated by bedload transport in rivers. The model provides an estimation of the acoustic power generated by impacts of bedload particles with riverbed particles. In this model, we account for bedload kinematics (e.g., impact velocities, saltation length) and the environment in which acoustic wave is propagated (e.g., acoustic wave attenuation). Sensitivity analysis shows that the noise generated by bedload transport depends on the grain size distribution and river characteristics such as slope, water level, and propagation effects. We tested the model on a field data set comprising acoustic and direct bedload measurements from different rivers. The acoustic powers predicted by the model are consistent with field measurements for some rivers while questionable for other ones. The analysis has shown a great sensitivity of the model to bedload kinematics (in particular the way the grain velocity is computed) and riverbed grain size distribution. The model provides a first basis that can serve as a framework in future work concerned with acoustic measurements of bedload transport. However, the model is still imperfect and it is limited by today's knowledge of the physics of bedload transport. Plain Language Summary Bedload refers to the transport of sediment particles on the riverbed. The study of bedload transport is important for flood hazard mitigation, protecting structures, and other diverse applications. A hydrophone is an instrument that records the acoustic noise generated by sediment transported on the riverbed, and can thus be used to monitor bedload transport. We developed a model to compute the acoustic noise generated by bedload transport to understand the important elements that influence the recorded noise. The results show that the model prediction errors vary significantly between different rivers. We found that the model errors are most likely due to errors in predicting transported particle trajectory and velocity. This lack of knowledge limits the performance of the model when predicting bedload transport flux with hydrophone measurements. These findings motivate new research perspectives.