The Hengduan Mountains are susceptible to hydrological disasters, with precipitation representing a significant risk factor. For effective disaster mitigation strategies, accurate rainfall simulation is essential, typically achieved through the use of numerical models. Some research has indicated that using a convection-permitting model (CPM) at high resolution (< 4 km) could provide more precise rainfall estimates than traditional cumulus parameterization schemes (CPs) at lower resolutions, but CPM demands substantial computational resources. Therefore, to assess whether CPM maintains superior simulation accuracy, this study employed the Weather Research and Forecasting (WRF) model to simulate summer precipitation over the Hengduan Mountains in 2009, comparing CPM (4 km) and CPs (10 km) resolutions. The simulations were evaluated against satellite observations to quantify their performance differences. The results showed that all simulations overestimated amounts and frequency. The CPM outperformed most CPs, except the Tiedtke scheme, which exhibited Root Mean Square Errors (RMSEs) of 2.51 mmday-1 for amount and 5.63% for frequency. The CPM had slightly higher RMSEs of 2.80 mmday-1 and 6.98%, respectively. Both CPM and Tiedtke captured the spatial distribution of precipitation, but overestimations occurred in central and southern regions and underestimations in river valleys. While Tiedtke demonstrated superiority in various aspects, CPM provided more detail. Additionally, the study noted significant differences in diurnal variation at intermediate altitudes and found correlations between rainfall amounts and convective available potential energy (CAPE), frequency, and outgoing longwave radiation (OLR), respectively. Consequently, the Tiedtke scheme is suggested as a more resource-efficient alternative to CPM for simulating precipitation in the Hengduan Mountains.