The escalating sophistication of cyber threats necessitates advanced and robust methods for malware detection. This study explores the effectiveness of machine learning models in identifying and classifying malware based on memory dumps, using the comprehensive CICMalMem-2022 dataset. We employed several machine learning techniques, including Random Forest, Support Vector Machine (SVM), Gradient Boosting, and Multi-layer Perceptron (MLP), to analyze and classify memory-based malware samples. Our results demonstrate that Random Forest outperformed other models, achieving the highest accuracy, precision, recall, and F1-score. Gradient Boosting also showed strong performance, particularly in adapting to complex and evolving malware behaviors. While SVM and MLP provided valuable insights into specific malware types, their overall metrics were slightly lower. This research underscores the potential of diverse machine learning approaches in enhancing malware detection systems. The findings contribute to the ongoing development of more sophisticated, real-time cybersecurity measures, paving the way for future work that integrates additional behavioral and heuristic indicators to further improve detection accuracy.