The long-term and excessive use of mineral fertilizers in a semi-arid region with severe water shortage will lead to soil compaction and poor water-holding capacity. The fertilization method of manure instead of mineral fertilizer has attracted wide attention. It has adverse consequences for the growth and development of crops. Hence, the objective of this study was to determine how replacing mineral fertilizer with manure affects the soil water retention curve, soil water constant, soil water availability, and soil equivalent pore size distribution, and to seek the best scheme of applying manure in semi-arid area and provide theoretical a basis for improving soil water retention capacity. Here, 0% (CK), 25% (M25), 50% (M50), 75% (M75), and 100% (M100) of 225 kg ha−1 nitrogen from mineral fertilizer were replaced with equivalent nitrogen from manure in the Loess Plateau of China under semi-arid conditions. The centrifuge method was used to determine the soil volumetric water content under different water suction levels, and the Gardner model was used to fit and draw its soil water retention curve, and then calculate the soil water constant and equivalent pore size distribution. The results showed that the Gardner model fitted well. The soil saturated water content with the M100 treatment was the highest, whereas the specific water capacity, water availability, and soil porosity with the M75 treatment were the highest. The soil saturated water content showed a downward trend with the increase in nitrogen from manure instead of nitrogen from mineral fertilizer in the partial replacement treatments. This downward trend slowed down over time. The M75 treatment increased field capacity. The M100 treatment increased soil capillary porosity, soil available water porosity, and soil water availability compared with CK from the fifth fertilization. Replacement treatments increased the specific water capacity, soil saturated water content, soil water availability, soil porosity, and reduced the wilting point over time. In the replacement treatments, specific soil water capacity, soil water availability, and soil porosity first rose and then declined with the increase in nitrogen provided by manure replacing that provided by mineral fertilizer. Therefore, the soil water holding capacity and water supply capacity with the M75 treatment were the best.