The BCC_AVIM model neglects the effects of molecular diffusion rate，vegetation litter and dry sur‐ face layer thickness on water vapour conductivity in the simulation of latent heat fluxes at the land surface. Based on this，this paper considers the above detailed physical processes for the simulation of latent heat fluxes at the land surface，and focuses on the effects of these processes on the simulation. The physical processes considered mainly include（1）the diffusion rate of molecules through the soil pores to the drying surface layer and the effect of plant apoplastic cover on the water vapour conductivity（Sakaguchi and Zeng，2009，hereafter referred to as the S-Z scheme）；（2）the effect of the thickness of the drying surface layer of soil on the water vapour conductivity（Swenson and Lawrence，2014，hereafter referred to as the S-L scheme）；and（3）the effect of physical processes on water vapour conductivity in the combined S-Z and S-L schemes，hereafter referred to as the SZ-SL scheme. In order to better study its sensitivity in different subsurfaces，observations from the flat and uniform al‐ pine meadow Maqu station and the complex topography of Si'e Mountain station in Southern Sichuan Forest were selected to examine the simulated effects of different parameterisation schemes. The results are as follows：（1） The BCC_AVIM model can better reproduce the trends of soil temperature and humidity，sensible latent heat fluxes and other elements at Si'e Mountain and Maqu stations，but the simulated values deviate from the measured values to a certain extent；（2）Based on the BCC_AVIM model，the simulated latent heat fluxes at the two stations are significantly improved by the introduction of more detailed physical processes，with the simulated results of the S-Z scheme being the closest to the measured values，and the simulated results of the S-Z scheme being the closest to the measured values. were closest to the measured values，with the Pearson's correlation coefficient（R）of Si'e Mountain station improved from 0. 68 to 0. 74，the mean absolute percentage error（MAPE）reduced by 10. 7%，and the root mean square error（RMSE）reduced by 4. 0 W·m-2；the R of Maqu station was improved from 0. 15 to 0. 62，the MAPE was reduced by 5. 7%，and the RMSE was reduced by 11. 2 W·m-2； （3）the simulation results of the two stations' soil temperature and humidity simulation results show that com‐ pared with the original scheme，the scheme after considering more detailed physical processes increases the simulated values of soil temperature and soil humidity and enhances the correlation with the measured values，of which the S-Z scheme is the optimal scheme for soil temperature simulation at the two stations，the S-L scheme is the optimal parameterization scheme for soil humidity simulation at Si'e Mountain station，and the SZ-SL scheme is the optimal parameterization scheme for soil moisture simulation at Maqu station. scheme；（4）by considering more detailed physical processes，not only the simulation accuracy of latent heat fluxes is significantly improved，but also the simulation ability of soil temperature and soil moisture is enhanced，but the improvement of sensible heat fluxes is still limited.