The traditional SWCC test method is time-consuming, and it is of great practical significance to develop a method that can quickly determine the SWCC of unsaturated soil. In order to predict the SWCC of compacted loess rapidly, the water potential and moisture content of compacted loess with different dry densities were tested, and the pore size distribution curve was measured using nuclear magnetic resonance (NMR) technology. Based on the test results, a rapid prediction method of the soil-water characteristic curve of Yan’an compacted loess was established based on the void ratio, and its accuracy was verified by the measured data. The results show that the fractal dimension D in the prediction model can be determined by the cumulative pore volume of two points (peak point and half-width point) on the pore size distribution curve and the slope of the pore size in a double logarithmic coordinate; and it can be expressed by the void ratio based on the linear relationship between void ratio and dominant pore diameter in logarithmic coordinates. The inlet value of SWCC is controlled by the diameter of macropore; the slope of the transition section is controlled by the volume of mesopore. There is a critical pore size in compacted loess, the residual water content is mainly controlled by the pore volume whose pore size is smaller than the critical pore size, and an empirical method is proposed to calculate the residual volume water content. Compared with traditional methods, the proposed method can save a lot of time in determining SWCC.