Flow in an open channel exerts a longitudinal shear force on the wetted perimeter of the channel flow cross section, resulting in the erosion of channel bed and side-walls. The boundary shear stress is one of the most common hydraulic aspect of flow in open channels, which has to be applied in practical applications such as water conveyance systems, sediment traort, and river morphology studies. Nevertheless, determination of this parameter is complicated, thereby analytical models are not able to resolve such problem solely. The influences of secondary currents, cross sectional shape, and non-uniformity of bed and side-walls roughness distribution around the wetted perimeter make this problem much more complicated with several unknowns involved, even for simple cases including straight and prismatic open channel flows. The interaction of these factors makes the prediction of boundary shear stress as a very difficult task. In addition to the total boundary shear stress, detaching of the bed and side-wall shear stresses is a key focus of a number of studies on open channel flows. To study a channel migration or to prevent bank erosion, the side-wall shear stress has to be determined. Whereas, bed shear stress is almost applied for studying the bed-form resistance and sediment traort. In the present study, mean and local distribution of boundary shear stresses in trapezoidal and rectangular open channels are investigated, applying three different methods including (i) a semi-analytical method, namely Guo and Julien method (GJM), (ii) an empirical correlation, based on nonlinear regression analysis on all the available and reliable experimental data to be applied in GJM, and (iii) an analytical method, namely Yang and Lim method (YLM). To obtain the delimitations of trapezoidal cross sections, applying the GJM method, the Schwarz-Christoffel transformation is employed and a unique relation has been achieved for different side-walls slope of 15 to 80 degree. Comparison of the present results with those of experimental data indicated that the YLM involves significant errors in determining the mean and local boundary shear stresses. These are partly due to the fact that YLM considers a linear form for the delimitations, regardless of the effects of secondary currents. The obtained errors are increased by increasing roughness of the boundaries. Empirical correction factors are obtained to improve the mean and local boundary shear stress correlations for smooth and rough open channel flows. The present study applies lumped empirical coefficients to the derived equations to consider the effects of secondary currents, based on a trial and error procedure, applying a multi-variable functions of geometric parameters. Results of the present semi-analytical model agree well with the experimental data, indicating that the present derived equations are good tools to estimate the mean and local boundary distribution in smooth and rough trapezoidal open channel flows. Finally plots of dimensionless mean boundary shear stresses in trapezoidal open channel flows of smooth or homogeneously roughened boundaries, are presented to be applied with design engineers. The derived semi-empirical relations apply for both sub- and super-critical flows in narrow to wide and shallow to deep trapezoidal open channels. Keywords: Boundary shear stress, conformal mapping, rough boundary, secondary currents, smooth boundary