Malware distribution through spam email infections represents a growing danger for users because spam emails have evolved into a major malware transmission channel. Deceptive emails attempt to trick users who are unaware of landing links or websites, which results in multiple dangerous outcomes, such as identity theft and deceit. Users experience disruptions due to spam content, mainly when such content consists of advertisements. This research examines how TF-IDF techniques extricate features from email content while performing multifaceted tests on spam detection through five machine learning (ML) methods including: Logistic Regression (LR), K-Nearest Neighbor (KNN), Decision Tree (DT), Random Forest (RF), and Support Vector Machine (SVM) for spam filtration and implementing multiple evaluation metrics encompassing accuracy, precision, recall and F1-score and ROC curve for model assessment. The implemented SVM system displayed superior performance compared to the alternative ML approach for detecting spam contents within email messages. The results of this work are hoped to serve as a benchmark for future works that will investigate more features at the syntactic and semantic levels.