Wireless sensor networks (WSNs), with their capacity to collect, analyze, and transfer data from distributed sensor nodes, have emerged as a revolutionary technological innovation. But these networks are open to a variety of security threats, which highlights the need for robust IDS [1] solutions. To achieve this goal, this paper attempts to present a novel ensemble learning-based IDS model that makes use of majority voting technique. First, we use a data augmentation approach called SMOTE to balance the class distribution of a dataset. Second, we lower computing costs by using PSO to discover the most valuable features. Third, we offer several base models and investigate how combining several techniques can result in a model that is more accurate. Several machine learning models are used in this instance, along with their combinations. Preprocessed dataset is fed to six classifiers, including support vector machine (SVM), K nearest neighbors (KNN), logistic regression (LR), Random Forest (RF), Naïve Bayes (NB), decision trees (DT). Our comprehensive research [2] identifies the optimal model for intrusion detection in wireless sensor networks (WSNs) using a huge dataset of 374,661 records from the wireless sensor network (WSN-DS) [3]. With an accuracy rate of 98.23% in binary classification settings and an impressive accuracy rate of 96.54% in multiclass classification settings, the suggested approach effectively identifies and mitigates intrusions in WSNs.