<p>Rapid and accurate diagnosis of the illness and related disorders is increasingly important to limiting the spread of the coronavirus disease, relaxing lockdown requirements, and reducing the burden on public health infrastructures. Recently, a variety of techniques and approaches have been proposed to classify the coronavirus using different clinical data and medical pictures. There are certain limitations and disadvantages to the coronavirus detection technology now available. Because of this, it is essential to develop and study new diagnostic tools that are more accurate while avoiding the shortcomings of existing tools. Using the SARS-CoV-2 CT scan dataset, this work separately evaluated non-linear SVM and Twin-SVM classifiers along with textural characteristics such as GLCM, GLRLM, and ILMFD. There are 1252 positive SARS-CoV-2 infection signs and 1230 negative ones among the 2482 CT scan images in this dataset. Eight different models were developed in this study to classify and predict the coronavirus. GLCM + NLSVM using RBF kernal, GLCM + TWSVM using linear kernal, GLRLM + NLSVM using RBF kernal, GLRLM + TWSVM using sigmoid, ILMFD + NLSVM using RBF kernal, ILMFD + TWSVM using polynomial kernal, Hybrid feature + NLSVM, and Hybrid feature + TWSVM were the models that outperformed when evaluated using the performance metrics used in this work. The Hybrid feature + NLSVM model with Linear Kernal yielded significantly better results than the other eight models tested for the dataset, including 100% accuracy, 100% recall, 100% precision, 100% F1-score, R-Squared = 1, and RMSE = 0. The speed and accuracy of the coronavirus diagnosis would therefore be greatly increased by the high accuracy of this kind of computer-aided screening approach, which would also promote the investigation of other related diseases using CT-scan pictures.</p>