Malware continues to pose a persistent and dynamic threat in the digital realm, demanding innovative detection and classification strategies. This study emphasizes the importance of feature fusion, which integrates diverse attributes from multiple sources to capture both static and dynamic aspects of malware comprehensively. Conventional single-feature approaches exhibit precision limitations, driving the exploration of multiple characteristics for fusion and the adoption of a unified learning algorithm for malware family classification. The research methodology involves meticulous feature extraction, followed by the application of KNN, XGBoost, DecisionTree, and random forest algorithms for classification, leveraging the most critical features. Experimental findings demonstrate a significant enhancement in classification accuracy compared to traditional methods, effectively reducing false positives. Fusion techniques enhance malware classification accuracy by 99.11% with dynamic features, 97.31% with static features, and 99.88% with hybrid analysis, surpassing conventional methodologies.