Background: Breast cancer patients are exposed to cardiovascular complications related to both pre-existing cardiovascular risk factors and cardiotoxic cancer therapies. Heart failure with preserved ejection fraction (HFpEF) is an emerging and often under-recognized condition in this population, potentially present before treatment initiation and further influenced by oncologic therapies. This study aimed to characterize baseline HFpEF probability and its longitudinal evolution using validated scoring systems in a real-world cardio-oncology cohort. Methods: We retrospectively evaluated 145 women with breast cancer undergoing baseline cardio-oncology assessment before initiation of potentially cardiotoxic therapies. HFpEF probability was estimated using three validated algorithms (H₂FPEF, ABA, and HFA-PEFF). A subgroup of 89 patients underwent exploratory reassessment after 12 ± 3 months to investigate longitudinal changes in HFpEF probability according to treatment. Results: At baseline, HFpEF probability varied substantially according to the scoring system applied, with 14.8–41.5% of patients classified as having intermediate-to-high probability. Agreement among the three algorithms was limited, particularly between HFA-PEFF and H₂FPEF/ABA scores. Exploratory longitudinal analyses showed no significant changes in HFpEF probability categories during follow-up. Conclusions: HFpEF probability assessment in breast cancer patients showed substantial baseline heterogeneity across validated scoring systems, with clinically relevant differences in risk classification and limited inter-score agreement. These findings provide novel insights into an underexplored cardio-oncology phenotype and suggest that currently available HFpEF scoring systems may not be directly interchangeable in this clinical setting. Prospective studies including provocative testing are needed to determine the clinical relevance of HFpEF probability scores in breast cancer patients.