<p>Rivers are among the primary sources for meeting drinking water, industrial, and agricultural needs worldwide, and from an environmental perspective, water quality in rivers is crucial for the survival of animal and plant species. Therefore, understanding and studying pollution transport processes in rivers is important. Analyzing the advection–dispersion equation, which governs pollutant behavior in surface flows, reveals that the longitudinal dispersion coefficient is the most critical parameter in this equation. Various methods have been proposed to estimate and calculate this coefficient. However, the range of variation in the longitudinal dispersion coefficient in natural rivers depends heavily on the hydraulic characteristics of the flow and the cross-sectional geometry, making accurate estimation of this complex parameter challenging. This study conducted a field investigation in two river sections to estimate the longitudinal dispersion coefficient based on reliable concentration-based methods and models that utilize the river's geometric and hydraulic characteristics. The accuracy and error of these methods in estimating the longitudinal dispersion coefficient were evaluated using statistical indices, including the Nash–Sutcliffe coefficient, average absolute relative error (AARE), and root mean square error (RMSE). The results indicated that the empirical models by Disley, Sattar and Gharabaghi and Noori performed best in estimating the longitudinal dispersion coefficient. Finally, an examination of the influential parameters showed that an increase in Manning's coefficient, shear velocity, and the U/u* parameter enhances the longitudinal dispersion coefficient in rivers. Consistent with previous research findings, the dimensionless parameter U/u* has a much greater effect on longitudinal dispersion in rivers than the W/H parameter.</p>