Disease diagnosis is conducted with a machine learning method. World Health Organization (WHO) reports that worldwide 7.6 million deaths are caused by cancer each year. Uncontrollable cell development in the tissues of the lung is called as lung cancer. These uncontrollable cells restrict the growth of healthy lung tissues. If not treated, this growth can spread beyond the lung in the nearby tissue called metastasis and, form tumours. In order to preserve the life of the people who are suffered by the lung cancer disease, it should be prediagnosed. So there is a need of pre diagnosis system for lung cancer disease which should provide enhanced result. We have proposed a novel machine learning method that hybridizes support vector machine (SVM), particle swarm optimization (PSO).The new method consists of two stages: initially, pso based approach for parameter optimization of SVM is developed to find the better initial parameters of kernel function, and then PSO is applied to continue SVM training and find the best parameters of SVM. Experimental results indicate that the proposed PSO-SVM model achieves better classification accuracy and F-measure than PSO-SVM and GASVM. Therefore, we can conclude that our proposed method is very efficient compared to the formerly reported algorithms.