Most research on the impact of medication reconciliation on patient safety focused on the retroactive model, with limited attention given to the proactive model. This study was conducted to compare the proactive and retroactive models in patients hospitalized for acute decompensated heart failure. This prospective, quasi-experimental study was conducted over six months, from June to November 2022, at the cardiology unit of an academic hospital in Iran. Eligible patients were those hospitalized for acute decompensated heart failure using a minimum of five regular medications before admission. Medication reconciliation was performed in 81 cases using the proactive model and in 81 using the retroactive model. 556 medications were reconciled using the retroactive model, and 581 were reconciled using the proactive model. In the retroactive cases, 341 discrepancies (both intentional and unintentional) were identified, compared to 231 in the proactive cases. The proportion of patients with at least one unintentional discrepancy was significantly lower in the proactive cases than in the retroactive cases (23.80% versus 74.03%). Moreover, the number of unintentional discrepancies was significantly lower in the proactive cases compared to the retroactive cases (22 out of 231 discrepancies versus 150 out of 341 discrepancies). In the retroactive cases, medication omission was the most frequent type of unintentional discrepancy (44.00). About, 42.70% of reconciliation errors detected in the retroactive cases were judged to have the potential to cause moderate to severe harm. While the average time spent obtaining medication history was similar in both models (00:27 [h: min] versus 00:30), the average time needed to complete the entire process was significantly shorter in the proactive model compared to the retroactive model (00:41 min versus 00:51). This study highlighted that the proactive model is a timely and effective method of medication reconciliation, particularly in improving medication safety for high-risk patients.