Aims: This study aimd to assess the association between estimated glomerular filtration rate (eGFR) levels and the prevalence of potentially inappropriate medications (PIMs) based on three criteria: the Beers (2019 version), Chinese and STOPP criteria. Methods: This retrospective cross-sectional study analysed 191 inpatients aged ≥65 years. Using the Beers, Chinese, and STOPP criteria, we assessed the prevalence of PIMs across stratified eGFR levels. Non-parametric analyses (Mann-Whitney U test), chi-square tests, and Logistic regression analysis were employed to evaluate the association between eGFR and PIMs exposure. Results: Participants were stratified into three groups based on eGFR: Group I (≥90 mL/min/1.73m²), Group II (60-89 mL/min/1.73m²), and Group III (<60 mL/min/1.73m²). Prevalence rates of PIMs were 72.25% by Beers criteria, 60.73% by Chinese criteria, and 28.80% by STOPP criteria, with STOPP demonstrating significantly lower PIMs detection rates than the other two criteria. Analysis by eGFR stratification revealed escalating PIMs prevalence with declining kidney function under Beers criteria: 67.35% in Group I, 74.29% in Group II, and 86.96% in Group III (P<0.05). Notably, Group III (eGFR <60 mL/min/1.73m²) exhibited higher proportions of patients with ≥2 PIMs across all criteria. By Beers criteria, 56.52% of Group III patients had ≥2 PIMs versus 22.45% (Group I) and 25.71% (Group II). Similar trends were observed by Chinese and STOPP criteria (P<0.05). Logistic regression analysis identified polypharmacy and eGFR decline as independent risk factors for PIMs exposure. Conclusion: The prevalence of PIMs gradually increases with the decline of eGFR in older adults, particularly when compounded by polypharmacy.