Soil erosion and sediment yield in basins are influenced by a combination of land use/land cover changes and climatic factors. The existing empirical Sediment Yield Index (SYI) model does not consider the spatiotemporal non-stationarity of the parameters and its application is limited in data deficient basins. To address these issues, a novel framework was proposed to extend the application of the empirical SYI model, and to identify the spatiotemporal heterogeneity of the driving factors by using the Geographically and Temporally Weighted Regression (GTWR) model and it was demonstrated in the Dongjiang River basin, South China. The constructed multi-factor GTWR model explains 87 % of the variation in SYI. Spatially, population density and urban land are the main driving factors of SYI, and there are significant spatial differences in their effects. Temporally, urban land is the most significant driving factor for SYI, and its effect increased over the year.