The IoMT is now a critical part of healthcare, data retrieval from medical devices and facilitating real-time monitoring. Class imbalance is one of the primary challenges related to IoMT datasets that in most cases minimizes the performance of IDS in properly identifying malicious behavior in healthcare networks. In this paper, an EL-based approach utilizing SMOTE to address the class imbalance problem in IoMT IDS. Instead of using the conventional parameter tuning, which can become inefficient in imbalanced data sets, our approach uses SMOTE-based oversampling schemes to create synthetic samples and balance class distribution. With the integration of SMOTE with EL method Boosting, the IDS enhances detection accuracy and stability. Results show that the SMOTE+Boosting classifier (SBC) method significantly enhances the classification performance relative to parameter tuning and feature se-lection regarding higher detection rates and less false positives. This article sheds light on the effectiveness of SMOTE-based oversampling in IoMT IDSs, particularly when handling highly imbalanced datasets.