Sediment settling velocities are commonly estimated from process-based or parametric data-driven approaches. The process-based approach has theoretical constraints due to the unclear settling physics; the parametric data-driven approach is limited by its mathematical assumptions. To overcome these limitations, this study compiles an aggregated sediment settling experimental database from literature and develops a non-parametric data-driven model to estimate the non-cohesive sediment settling velocity in water. A cross-comparison against five process-based equations and a parametric data-driven equation demonstrates the higher accuracy and better consistency of the new model in estimating sediment settling velocities under various physical regimes. The data-driven model also shows an easy-implemented self-update capability by assimilating theoretical data generated from the process-based equations. The updated model, leveraging experimental and theoretical data of sediment settling process, further improves the accuracy and reduces the uncertainty in estimating sediment settling velocities. This approach illustrates the value of integrating experimental and theoretical knowledge in estimating the complex process in sediment transport, and provides an alternative framework for future sediment transport exploration.