【Objective】Maintaining moderate soil phosphorus (P) supply intensity and determining the P application rate based on the P requirement of the crop target yield, and combining with acid, slow-release water-soluble P fertilizers and applied in the early stages of crop growth is a novel P management approach in cotton. Here we test the feasibility and effectiveness of this method and provide a basis for reducing the P fertilizer application amount and increasing its use efficiency. 【Method】A two-year field experiment was employed using cotton (Gossypium hirsutum) in the moderate Olsen-P soil (23.6 mg?kg-1) in Shihezi, Xinjiang. Based on the P requirement of the target lint yield of 2.7 t?hm-2, 52.4 kg?hm-2 was selected as the P application rate. Four treatments were set-up with three types P fertilizers: (1) no P fertilizer (CK); (2) monoammonium phosphate (MAP); (3) urea phosphate (UP); and (4) ammonium polyphosphate (APP). It was analyzed that the lint yield, P uptake and above-ground biomass of cotton in different treatments. The partial productivity of P fertilizer (PFP), agronomic efficiency (AEP), P utilization efficiency (PUE) and P recovery index (the ratio of shoot P uptake to the amount of P fertilizer application, PRI) were calculated. Through stratified sampling, soil Olsen-P and P fractions in different soil layers were tested. Those allowed us to clarify the relationships between the P spatial availability and cotton growth requirements with different P fertilizers. 【Result】The significant results were: (1) The average lint yield of P applied treatments in two years was 2.73 t?hm-2, which was classified as the high yield level in northern Xinjiang. Compared with the CK treatment, the lint yield of MAP, UP, and APP treatments were increased by 21%, 25%, and 11%, respectively. (2) The P surplus was 2.2-4.8 kg?hm-2 and the P use efficiency was 24% across P applied treatments. The P recovery index of MAP, UP, and APP treatments was 91.7%, 95.6%, and 94.8%, respectively. (3) P fertilizer applied by drip irrigation in the cotton bud and boll stages can move to the 10-20 cm soil layer, which may maintain the soil P supply intensity during the cotton-growing period. Compared with the CK treatment, the mean Olsen-P in P applied treatments were increased by 94%~302%, 104%~144%, and 42%~67% in the 0~5 cm, 5~10 cm, and 10~20 cm soil layers, respectively. 【Conclusion】In summary, our results indicate that maintaining soil available P near the agronomic thresholds and determining the P application rate according to the P requirement of the target yield, and combining with acid, slow-release P fertilizer and applied in the cotton bud and boll stages can improve lint yield and P use efficacy under balanced P input-output conditions. Furthermore, the P recovery index based on the balanced P input-output can better reflect the real P use efficiency.