This study introduces an enhanced soft-computing-based strategy for efficient feature selection in breast cancer (BC) prediction using the Wisconsin Diagnostic Breast Cancer (WDBC) dataset. The authors propose three feature selection methods: Gravitational Search Optimization Algorithm (GSA), Emperor Penguin Optimization (EPO), and a hybrid approach called hGSAEPO. These methods aim to identify the most relevant features while reducing irrelevant ones, thereby simplifying complexity and improving accuracy. The study focuses on BC due to its high mortality rate among women, emphasizing the importance of early detection and diagnosis. The experimental results, involving eight tests, demonstrate that the proposed approach achieves excellent performance in binary classification, with precision, sensitivity, specificity, F1-score, and AUC values exceeding 0.95. This robust clinical prediction system is designed to assist healthcare professionals in making informed decisions about BC diagnosis and treatment.This study introduces an enhanced soft-computing-based strategy for efficient feature selection in breast cancer (BC) prediction using the Wisconsin Diagnostic Breast Cancer (WDBC) dataset. The authors propose three feature selection methods: Gravitational Search Optimization Algorithm (GSA), Emperor Penguin Optimization (EPO), and a hybrid approach called hGSAEPO. These methods aim to identify the most relevant features while reducing irrelevant ones, thereby simplifying complexity and improving accuracy. The study focuses on BC due to its high mortality rate among women, emphasizing the importance of early detection and diagnosis. The experimental results, involving eight tests, demonstrate that the proposed approach achieves excellent performance in binary classification, with precision, sensitivity, specificity, F1-score, and AUC values exceeding 0.95. This robust clinical prediction system is designed to assist healthcare professionals in making informed decisions about BC diagnosis and treatment.