Background Older adults in residential aged care (RAC) experience a high burden of medication-related problems, increasing the risk of unplanned hospitalisation. A medication-based risk tool may offer a more feasible and scalable alternative in settings with limited access to high-quality clinical data. Objectives To develop and internally validate the Medication-related Hospitalisation Risk Index (MeHRI), a medication-profile–based tool for predicting unplanned hospitalisation risk in RAC. Methods A retrospective cohort study using electronic health records from 24 RAC homes in Sydney, Australia (2014–2019). The outcome was time to first unplanned hospitalisation within 12 months of entry. Cox proportional hazards models were used to identify predictors, and a regression coefficient–based method was applied to derive a point-based risk score. Internal validation was performed using 5-fold cross-validation, and discrimination assessed using AUROC. Results The study included 2,417 residents; 905 (37.4%) experienced unplanned hospitalisation within 12 months. The final model included 19 medication-related variables. Increased risk was associated with medications for gastrointestinal, cardiovascular, infectious, respiratory, pain, epilepsy indications, systemic corticosteroids, and higher medication count. Anti-dementia therapy was inversely associated with hospitalisation (HR 0.53, 95% CI 0.41–0.70). The MeHRI score ranged from −4 to 38 and was categorised as low (<3), medium (3–9), and high (≥10) risk. Model discrimination was moderate (AUROC 0.65). Conclusion MeHRI is a pragmatic medication-based risk index for identifying RAC residents at increased risk of unplanned hospitalisation. It enables routine risk stratification using readily available medication data. External validation and assessment of clinical utility are warranted.