Due to the expanded use of the internet and the revolution of the information technology field, people are beginning to read news online more and more. For that reason, online news has become the main source of information for the majority of people, and predicting the popularity of online news has become a hot topic as it could help writers present competitive and highly readable news. These predictions can be done by using machine learning techniques. This paper introduces some of the most well-known prediction classification models in data mining like Random Forest, Bayes Net, Logistic Function, C4.5 and Simple Cart which has been applied in order to predict the popularity of the online news. The objective of this paper is to evaluate the performance of these different models on real-world online news data. The experiment results revealed the success of some of these models in predicting the popularity of online news with relatively high accuracy. The performance of the five models is evaluated by some of the most popular metrics such as Accuracy, Root Mean Square Error (RMSE), Kappa Statistic, TP-Rate, FP-Rate, Precision, F-Measure and ROC Area values. Finally, feature filtering techniques are applied in the study to improve the model's performance and identify the most influential features affecting news popularity.