Early prediction of treatment response in eating disorders remains challenging, and the platelet-to-lymphocyte ratio (PLR) is a readily available inflammatory biomarker. This study aims to investigate the association between PLR and early treatment response in patients with eating disorders. We retrospectively examined whether the PLR, a routine inflammatory index, tracked with early improvement in 64 inpatients treated at Shandong Daizhuang Hospital from June 2013 to December 2023. Response was defined using the Clinical Global Impression-Improvement scale after two weeks of treatment. Twenty-eight patients (43.8%) responded. Responders had lower baseline PLR (111.62 ± 36.50 vs. 151.97 ± 67.38, Z = 2.287, p = 0.022) and shorter illness duration (Z = 2.000, p = 0.046). In a multivariable model adjusting for duration, PLR was marginally associated with response (odds ratio = 0.987, 95% CI: 0.9747–0.99996, p = 0.049); because covariates were selected based on univariate significance rather than prespecified clinical criteria, this estimate should not be read as an independently adjusted effect. On the receiver operating characteristic analysis, the area under the curve for PLR was 0.668 (95% CI: 0.535–0.801), with 82.1% sensitivity and 55.6% specificity at a cutoff of 137.75. The signal is real but weak: the CI barely excludes the null, the result would not survive multiple-testing correction, and specificity is modest. PLR is thus unlikely to stand alone in clinical decision-making, though it may serve as a low-cost adjunct. These preliminary findings require confirmation in larger prospective cohorts.