A total differential equation was proposed to assess the driving factors for the spatial variations in potential evapotranspiration (ET0). Using China's Loess Plateau as an example study area, three transects with distinct ET0 gradients in space, i.e., northwest–east, northwest–south and northwest–southwest, were chosen to sample spatially varied ET0 and four climatic variables (solar radiation, actual vapor pressure, wind speed, and mean temperature) at an interval of 10 km. Considered an independent variable, the distance was differentiated to quantify the contribution of each climatic variable to the spatial ET0 variations along each transect. A significant decrease in solar radiation and an increase in actual vapor pressure were identified as the dominant impact factors that led to a decreased ET0 in the northwest–east and northwest–south directions, respectively. As another key contributor, the decreasing wind speed induced a decreasing trend in ET0 from northwest to southwest. The above results implied that the dominant factor(s) for the spatial variations in ET0 differed among the regions. Therefore, the total differential equation is a powerful approach to determine the driving factors and to quantify their individual contribution to the spatial variations in ET0.