The advent of artificial intelligence (AI) has brought about a significant transformation in breast cancer (BC) prediction, resulting in faster and more precise identification of cancer. Gene expression (GE) data plays a crucial role in identifying biomarkers, classifying molecular subtypes, predicting therapy responses, and uncovering therapeutic targets, all of which contribute to tailoring treatment approaches. Next-generation sequencing (NGS) outperforms microarray gene expression (GE) analysis by providing superior resolution, comprehensive assessment, and improved accuracy, thereby boosting the evaluation of risk, selection of medication, and prediction of prognosis. This research examines different machine learning (ML) methods for identifying breast cancer using NGS-GE data, with a specific emphasis on minimum redundancy maximum relevance (MRMR) for selecting features. The proposed MrmrXgBr model utilizes three widely-used ML algorithms: eXtreme gradient boosting (XGBoost), support vector machine (SVM), and random forest (RF). The selection of these algorithms was based on their strong classification performance, capability to handle data with a large number of dimensions, ability to evaluate feature importance, and capacity to identify intricate connections in GE datasets. The experimental findings showcase our proposed MrmrXgBr model’s efficacy, with an impressive 96% accuracy in classification, 100% sensitivity, and 97% F1-score. These findings highlight the significance of accurate feature selection in improving the efficacy of ML models for BC detection. This research greatly enhances the application of advanced ML techniques and efficient feature engineering in creating more robust and effective tools for identifying BC. As a result, it improves patient outcomes and assists in making therapeutic decisions.