The creation of efficient fake news detection systems has become necessary due to the digital age’s quick spread of false information. This research delves into the creation and evaluation of a machine learning model that geared toward identifying false information in news articles. Capitalizing on a “Fake News Detection” dataset that has been acquired from Kaggle, that comprises of labeled examples of real and fake news, this study accentuates the importance of feature extraction techniques and model selection. An assortment of machine learning algorithms has been employed, including Support Vector Machines (SVM), Logistic Regression, Naive Bayes, Random Forest, Decision Tree, Gradient Boosting Machines (GBMs), Light GBM, XG Boost, Singular Value Decomposition (SVD), alongside a voting ensemble model, each evaluated based on their predictive accuracy and performance metrics. Among the models tested, the SVM achieved the highest accuracy at 98.12%, followed closely by both Light GBM and SVD with 97.99%.Using the strengths of SVM, Random Forest, and XGBoost classifiers, the ensemble produces predictions that are 98.75% accurate. The voting ensemble achieved impressive performance, with the combined accuracy surpassing individual models. The study also highlights the significance of hybrid feature extraction methods, including TF-IDF, N-grams, and Word2Vec, in enhancing model performance. Through rigorous evaluation, this research demonstrates the potential for machine learning models to effectively combat the spread of misinformation, providing implications for future advancements in automated fact-checking systems.