Aim:The main aim of the study is to improve junk email detection by using machine learning algorithms with Novel Support Vector Machines and Naive Bayes. Materials and Methods: In this work, we employed both new support vector machines and Naive Bayes to investigate their effectiveness in detecting spam emails. Using the G Power program, we calculated the sample size to be 10 in group, with pertest power of 2, The threshold of 50 & confidence interval of 95%. The results showed that the Novel Support Vector Machine outperformed the Naive Bayes method in detecting email spam, with an average accuracy of 93.52% compared to 82.35% for Naive Bayes. Significance value of a p = 0.027 indicates a significant difference between the two groups (p < 0.05). Conclusion: In the conclusion, our findings suggest that Novel SVM Machine approach is more effective than Naive Bayes method in detecting spam emails.