Straw return can improve soil nutrient availability, but its interaction with phosphorus (P) fertilizer on cotton (Gossypium hirsutum L.) production and soil P availability has rarely been studied. Therefore, we carried out a four-year field trial to evaluate the effects of straw management (retention or removal) with P fertilizer (0, 22, 44, 66, 88 kg P ha–1) additions on seed cotton yield, cotton P uptake, soil fertility, and P apparent recovery efficiency (PRE) in a continuous barley–cotton rotation system in a coastal saline soil field from 2016-2019. Our results showed that the year and the interaction of P fertilization and straw management had significant effects on seed cotton yield and boll numbers, but there was no significant effects of straw management and the interaction of P fertilization and straw management on seed cotton weight per boll. An increase in P activation coefficient (14-28%) was observed due to the higher soil available P content following straw retention compared to the straw removal. Compared with straw removal, straw retention reduced mineral P fertilization by about 27-38% without affecting seed cotton yield, while increasing PRE by 14‒27% over the four years. In the 0-20 cm soil layer, straw retention reduced the Langmuir P sorption maximum (Qmax) and maximum buffer capacity (BCmax), and also increased the degree of P saturation (DPS), but it did no effect on the soil adsorption equilibrium constant (k). Straw retention significantly decreased the soil Qmax compared to straw removal under 0, 22, and 44 kg P ha–1, while no significant straw effects were found on Qmax under 66 and 88 kg P ha–1. Cotton P uptake and seed cotton yield were negatively correlated with the Qmax (P<0.01). Our results suggest that straw return may be a promising management practice to reduce mineral P fertilization without significant loss of seed cotton yield on coastal saline lands.