The world has always suffered and from diseases and epidemics, and the coronavirus is one of the most dangerous viruses that threatened human life that requires the use of all scientific methods and means to respond to it and reduce its spread by early detection of infections and taking necessary measures In view of the significant role that artificial intelligence plays in most fields of science, it has become one of the most important scientific methods used to resolve complex issues and has been harnessed in medical diagnosis, one of the most complex areas. Many AI and machine learning algorithms have been used to diagnose and detect diseases in general and coronavirus in particular. The support vector machine (svm) machine algorithm was one of the most important algorithms in this area and is one of the most effective compilations used in the knowledge extraction process In spite of all this, the results they present remain incomplete because classification issues do not deal with cognitive uncertainties such as ambiguity, neutrality and inconsistency associated with perception of human thinking, This adversely affects the work of a classic support vector machine and affects the accurate diagnosis of the disease To solve this problem, we have done this research using a Neutrosophic Support Vector Machine because it takes into account all possible cases during the study of the sample and it reduces the impact of extreme values. This increases the accuracy of the results when diagnosing coronavirus symptoms. The study was conducted according to the following steps: 1. We extract features from chest radiographs based on GLCM 2. We form a neutrosophic dataset. 3. We train Neutrosophic Support Machine N-SVM on new data. 4. We record the results. Comparing the results, we got using the upgraded N-SVM algorithm with the classic SVM algorithm results we found that it gives a more accurate diagnosis of the disease.