The utilization efficiency of phosphorus fertilizer is generally low in agro-ecosystems as phosphorus is readily co-precipitated with iron and manganese oxyhydroxides, iron aluminides, or directly absorbed to the soil minerals, particularly in the red soil ecosystem. Thus, understanding the subsequent effects of phosphorus retained in the soil on soil ecological function and crop nutrient supply is of significance to guide the management of phosphorus fertilization. Based on a long-term fertilization experiment in red upland soil, this study investigated the effect of the conventional fertilization (CK) and short-term phosphorus fertilizer input (at rates of 0, 50, 100, 150 and 1 000 kg·hm-2, P2O5) on soil nutrients, soil nitrogen cycle process and crop yields after 27 years. We evaluated the relationship between these factors and the residual effect of phosphorus using the multivariate statistical analysis method. Short-term P fertilizer addition at a high rate (1 000 kg·hm-2, P1000) had no significant effects on soil total carbon (TC), total nitrogen (TN) and microbial biomass (MBC) compared to the conventional fertilization treatment (CK). Also, it significantly increased soil pH, nitrogen mineralization rate (Nmin), soil nitrogenase activity (SNA), and potential nitrification rate (PNR) (P < 0.05), while it significantly decreased the net N2O emission potential (NN2O) (P < 0.05). Compared with the low phosphorus dosage treatment (50, 100, 150 kg·hm-2), P1000 significantly increased soil available phosphorus (AP), Nmin, SNA, PNR and potential N2O production rate (PN2O) by 33.3%-76.4%, 88.2%-388.1%, 111.4%-4 826.3%, 22.6%-152.4% and 13.8%-78.9%, respectively (P < 0.05), but significantly decreased the net N2O emission potential by 64.6%-78.9% (P < 0.05). These results suggest that the application of a high dosage of phosphorus fertilizer has a strong residual effect on soil microbial activity and nitrogen processes even after 27 years. Spearman correlation analysis and redundancy analysis showed that AP and pH were the most important factors affecting soil microbial activities. Maize yield in the recent three years showed no significant difference among all treatments but was significantly positively correlated with TP, AP and pH. In comparison to low phosphorus treatments, P1000 treatments showed a promotion effect of 3%-23% on maize yield based on the historical yield data during 1991—2019. Our results reveal that the short-term application of a large amount of phosphorus fertilizer has significant positive effects on maintaining soil fertility, microbial activity, soil nitrogen cycle function activity and crop yield even after 27 years, owing to the promotion of soil pH and slow release of available phosphorus