Abstract: Excessive application of phosphate fertilizer wastes phosphorus resources and induces eutrophication in lakes and rivers. To study the effect of reduction of phosphorus fertilizer on phosphorus fractions in calcareous soil and its relationship with crop phosphorus accumulation, three treatments were set up, i.e., phosphorus application rates of 150 kg∙hm⁻² (P150), 37.5 kg·hm⁻² (P37), and 0 kg∙hm⁻² (P0). After two consecutive years of “winter wheat-summer maize” crops rotation, the changes in the contents of soil phosphorus fractions were studied using Hedley soil phosphorus fractionation method, and the storage contribution rate and output contribution rate of each fraction were also estimated. The relationship between soil phosphorus fractions contents, phosphorus fertilizer application rate, and crop phosphorus uptake amount were explored by using regression analysis, path analysis, and structural equation model. The results showed that compared with P37, P150 led to a significant increase in soil total phosphorus content. The contents of inorganic phosphorus extracted with anion exchangeresin (resin_P_i), with NaHCO₃ (NaHCO₃_P_i), with NH₄OAc (NH₄OAc_P_i) and with NaOH-Na₂S₂O₆ (Fe_P_i), and organic phosphorus extracted with NaHCO₃ (NaHCO₃_P_o) in P150 were significantly higher than those in P37, while the other fractions showed no significant change. P0 did not cause a significant decrease in the contents of soil phosphorus fractions. The storage contribution rates of soil inorganic phosphorus fractions and organic phosphorus fractions were 72.6% and 23.8%, respectively. Among them, the storage contribution rates of inorganic phosphorus extracted with HCl (HCl_P_i), Fe_P_i, NH₄OAc_P_i, resin_P_i, and organic phosphorus extracted with HCl (HCl_P_o) were 24.45%, 18.1%, 13.62%, 11.15%, and 9.30%, respectively. The output contribution rate of soil inorganic phosphorus fractions was 41.0%, and that of organic phosphorus fractions was 56.4%. Among them, the output contribution rates of HCl_P_o, Fe_P_i, and NH₄OAc_P_i were 39.44%, 17.36%, and 13.06%, respectively. The output contribution rates of HCl_P_i and resin_P_i were only 1.91% and 0.40%, respectively. In the structural equation model, the load factors of phosphorus fertilizer application rate on Fe_P_i, HCl_P_i, NH₄OAc_P_i, resin_P_i, organic phosphorus extracted with NH₄F (NH₄F_P_o), NaHCO₃_P_i, and NaHCO₃_P_o were 0.078, 0.077, 0.061, 0.036, 0.018, 0.015, and 0.012, respectively. The load factors of Fe_P_i, NH₄OAc_P_i, and HCl_P_o on crop phosphorus uptake were 0.355, 0.334, and −0.039, respectively. The above results show that in calcareous soil, Fe_P_i, NH₄OAc_P_i, and HCl_P_o were the key phosphorus fractions. Among them, Fe_P_i and NH₄OAc_P_i were easily consumed when no phosphorus fertilizer was applied, but they can be easily supplemented by phosphorus fertilizer application. However, HCl_P_o was available to the crop but was not easily replenished by phosphorus fertilizer application. The high storage contribution rate and low output contribution rate of HCl_P_i fraction were the important reasons for the low efficiency of phosphate fertilizer in the current season. It is suggested that the choice of phosphorus application rate should be based on the storage contribution rate of the key phosphorus fractions.