To estimate the leaf area index (LAI) in large areas, this paper analyzes the relationships between normalized difference vegetation index (NDVI) and the grassland LAI based on MODIS data in the southern grassy mountains and slopes of China. By using nonlinear fitting equation we constructed the basic estimation model of grassland LAI with NDVI as the independent variable and introduced precipitation and temperature as regulatory factors. The model was validated with observed data in different years and the results showed that there was a good correlation between the simulated and observed LAI value with a statistically significant level of R2. RMSE was 0.302 and RRMSE was 0.154. It was also found that the spatial distribution of grassland LAI in south China showed a remarkable zonal characterization, and temporal distribution showed a single peak curve. These results provided a theoretical basis for the effective management of southern grassland resources and the carbon sink estimation of the nationwide grasslands.