The method users five well-known algorithms to induce the spam classifiers: Random Forest (RF), BAGGING, ADABOOSTM1, Support Vector Machine (SVM), and Naïve Bayes (NB). Evaluating effectiveness of the classifier found that BAGGING performs the best. Moreover, the performance surpasses a number of state-of-the-art methods proposed in previous studies. Although applied only to English language e-mails, the results indicate that this method may be an excellent means to classify spam e-mails in other languages as well. The reason of choosing algorithms Random Forest (RF) is the spam detection has high accuracy and learns quickly, the operational algorithm is suitable for a large number of data.The secondary algorithm as an ensemble learning method is called ADABOOSTM1. Although an ensemble method, ADABOOSTM1 is both simple and fast. The biggest advantage of this ensemble method is that is less susceptible to training data over fit. The algorithm has been reported perform better then Naïve Bayes (NB) and Probabilistic TF-IDF for text categorization tasks. Support Vector Machine (SVM) is also a popular learning algorithm for spam detection, but is evident that the performance of SVM is better with header features than with text features. Naïve Bayes (NB) is a widely-used learning algorithm. In the anti-spamming community use NB to generate classifiers because the algorithm is simple yet powerful enough to detect spams effectively