The scale-aware physical process is a trend of modern numerical weather prediction. Considering that GRAPES_Meso (Global/Regional Assimilation and Prediction Enhanced System) model didn’t include this physical process, this paper introduces the scale-aware parameterization in the KFeta (Kain-Fritsch Eta) convective parameterization scheme firstly. The convective time scale, grid vertical velocity and entrainment rate of the scheme are improved based on the scale-aware parameterization. In order to analysis the impact on the results simulated by the scale-aware scheme and the original scheme in different resolution models, a squall line process in South China was selected. The results show that in the GRAPES_Meso model with horizontal resolution of 3km, 5km, 10km, and 20km, the scale-aware KFeta scheme has certain positive effects on the simulation of precipitation intensity and location. As the model resolution increasing, the sub-grid precipitation decreased and grid precipitation increased. The entrainment in the tropospheric middle and low layer slightly enhanced, the existing cold deviations have some improvement. The intensity of updraft and the hydrometeors content in the strong convection area are more consistent with the real atmospheric conditions. Overall, the modified scheme is more suitable for high resolution numerical weather prediction model. The results can provide useful reference for the application of scale-aware convective parameterization scheme and the optimization of strong precipitation forecasting performance in numerical weather prediction model.