Abstract: In order to study the characteristics of errors associated with the lateral boundary conditions used in regional modeling, especially the regional climate modeling, based on the YH model, a regional model system nested with a spherical-belt type model is established and a set of error diagnosis tools are developed. By means of the software Excel, a series of comparative studies are carried out to analyze the spatial and temporal distribution of the errors due to the different lateral boundary conditions. The results show that the basic methodology for detecting the errors associated with lateral boundary conditions in regional modeling is rational, and the tools are quite useful. Moreover, it might also be applied to other studies, such as for diagnosing other types of errors in the model or developing some new schemes. The preliminary results indicated that the lateral boundary error of a regional model varies with the location in the model domain, and the spatial and temporal distribution of the error associated with different variables is rather different. Relatively, the errors of mass and potential vorticity fields are smaller in the YH model. The errors of kinetic energy and water vapor field are found mainly in the influent part of upper layers in the near-boundary region, while the error of sensitive heat field was in the upper layers of near-boundary region. The fixed lateral boundary conditions cause more serious errors than the nested ones, and the time interval updating the lateral boundary also has some impact on the error. It is interesting to mention that the errors of upper layers in the fine mesh regional model where the long wave motion dominates are larger than the errors found in the coarse mesh spherical-belt one. On the other hand, the errors of lower layers in the same fine mesh regional model where the short wave motion dominates are smaller than the errors found in the coarse mesh spherical-belt one.The impact of phase velocity error on lateral boundary errors is more significant in the lower layer, which is related to when the ratio of grid lengths of fine-mesh vis-à-vis coarse-mesh. The errors became much more serious when the diurnal variation cycle is turned off in the regional model.