Abstract: Biochar can be utilized as an adsorption material to recover phosphorus (P) from wastewater, and then applied as the phosphorus fertilizer in soil for agricultural production. The simple, effective, and cost-effective way can be expected as the potential solution to environmental pollution and phosphorus resource scarcity. However, single biochar (BC) has relatively limited adsorption capacity for anions, due to its negative charge. Additionally, the functional groups can be lost in the processes of BC production during high-temperature pyrolysis. Therefore, more research tends to focus on modified BC and composite materials, in terms of their adsorption performance. Iron oxides, especially natural iron minerals, are widely used for P recovery, due to their abundant reserves, low cost, and fast reactivity. Loading hematite (H) onto the inner and outer surfaces of larger-sized BC can enhance the selective adsorption of BC for PO₄³⁻. Significant adsorbent loss can also be avoided in recovery when using H alone. In this study, the preparation process of modified walnut shell BC with H was optimized using a single-factor combined response surface method (RSM). Model fitting was used to analyze the adsorption characteristics of modified biochar (H-BC) for P in water. Finally, the modified biochar saturated with adsorbed phosphorus (H-BC-P) was applied to the soil in pot experiments. The performance of adsorption was verified to evaluate the potential of the P fertilizer. The results showed that the modified biochar produced by pyrolyzing at 850°C for 45 minutes achieved a P removal rate of 98.42% and a saturated adsorption capacity of 15.59 mg/g at a walnut shell/hematite mass ratio of 1:1. The adsorption process of P by H-BC was better described by the pseudo-second-order kinetics and the Freundlich model, indicating multilayer adsorption, with the rate-limiting step of chemical adsorption. Acidic conditions favored the adsorption of P by H-BC. The optimal dosage of H-BC was 2.5 g/L from a cost perspective, resulting in a P removal rate of 95.31%. BET test results showed that the H-BC had a specific surface area about 5.7 times that of BC, reaching 45.75 m²/g. There was a significant increase in the pore volume and average pore size. Therefore, the H modification posed a significant impact on the physical properties of BC. FT-IR, XRD, and EDS test results indicated that Fe₂O₃ in hematite was reduced to magnetite during co-pyrolysis, and a large amount of iron elements was loaded onto the surface of H-BC. SEM tests showed that the surface of H-BC-P exhibited numerous cluster-like substances, possibly FePO₄. When the H-BC-P was applied to the soil, there was an increase in the total P and Olsen-P content, as well as the pH of acidic red soil. But there was no effect on the pH of calcareous brown soil. The application of H-BC-P resulted in a significant increase in the lettuce biomass in both types of soil. In the calcareous brown soil, the fresh weight and leaf count of lettuce increased from 30.5% and 10% to 81.4% and 50%, respectively. In the red soil, there was an increase in the fresh weight and leaf count of lettuce from 50.9% and 18% to 101.8% and 36%, respectively. The magnetite-modified biochar was an environmentally friendly P adsorbent. The residue after P adsorption can be expected to serve as a P fertilizer for the potential to the acidic soil.