The method of support vector machines, also called large margin machines (SVM), is a derivative of computer learning, a solid theoretical basis different from its ancestor of neural networks. The SVM method is known as a very interesting advance in its principle, its implementation and its extension to multiclass problems. Over the past decade, research has focused on adapting the method to particular problems such as novelty detection and clustering as well as improving its performance (optimization and parallelization) and its application. in different domains. such as imaging, sound, banks, biology, management and knowledge management. As a result, several applications have experienced real success with the use of SVMs. In this article, we have highlighted in depth the correlation of SVMs especially on knowledge management as well as the companies implemented Corporate Social Responsibility (CRS). The purpose of the classification is to apply the variables according to the proposed method based on SVM improves the classification accuracy of the classifiers compared to the simple term-based data classification method in order to simplify and facilitate the decision making .