In this study, the effect of horizontal and vertical resolution on the precipitation forecast has been investigated. For this purpose, in this study, heavy rainfall occurred in the western region of Iran using GFS data as boundary and initial conditions in the WRF mesoscale numerical forecasting model with two horizontal resolutions and four combinations of position and number of vertical levels has been simulated. The results of this study show that on days when the intensity of precipitation is significant, the results of precipitation simulation improve with increasing horizontal resolution. Also, increasing the resolution of the model in the vertical direction does not necessarily lead to an increase in precipitation forecasting skills and sometimes over-predicted rainfall. Skills of forecasting improved for all rainfall thresholds when the resolution above the melting level was increased. And this is because of better prediction microphysical processes. Also, increasing the resolution of the model in the lower layer of troposphere only in heavy rainfall leads to an increase in forecasting skills. Therefore, the effect of increasing the resolution of the model on forecasting skills is strongly related to the height of added levels and rainfall intensity. In addition, the factor separation methodology shows that with increasing the resolution of the model throughout the troposphere layer, the prediction skill generally decreases in the mean state, and this subject is due to the negative interaction between lower-tropospheric processes and microphysical processes above melting level.