Objective To investigate the predictive value of fibroblast growth factor-23 (FGF-23) for postoperative acute kidney injury (AKI) in elderly patients undergoing abdominal surgery. Methods The medical records of elderly patients who underwent elective non-vascular intra-abdominal surgery at Shanxi Provincial People's Hospital from November 2022 to November 2023 were collected. Venous blood samples were obtained one day before surgery and one day after surgery, and serum FGF-23 levels were measured using enzyme-linked immunosorbent assay (ELISA). Serum creatinine (Scr) levels were monitored within seven days postoperatively, and patients were divided into an AKI group and a non-AKI (NAKI) group based on the results. Due to a significant imbalance in sample sizes between the two groups, propensity score matching (1:2) was applied to reclassify the groups. Statistical methods were used to analyze whether the differences in FGF-23 levels between the AKI and NAKI groups were significant. Results A total of 100 patients were included in this study, of whom 12 (12%) developed AKI. After propensity score matching (1:2), compared with the NAKI group, the AKI group had significantly higher preoperative day 1 serum FGF-23 levels, postoperative day 1 Scr levels, and postoperative day 1 serum FGF-23 levels (P 0.05). Multivariate logistic regression analysis showed that, after adjusting for confounding factors, preoperative day 1 serum FGF-23 levels and postoperative day 1 serum FGF-23 levels were independent risk factors for AKI in elderly abdominal surgery patients. ROC curve analysis revealed that preoperative day 1 serum FGF-23 (AUC = 0.868) had better predictive value than preoperative day 1 Scr (AUC = 0.716), while postoperative day 1 Scr (AUC = 0.926) had better predictive value than postoperative day 1 serum FGF-23 (AUC = 0.863). Conclusions Our findings suggest that preoperative serum FGF-23 levels are significantly associated with an increased risk of postoperative AKI in elderly abdominal surgery patients, indicating its potential value as an auxiliary predictive biomarker.