The apparition of Covid19 represents a horrible disease that upset the human's life. The difficulty of this disease is its rapid evolution through people contact. Hence, designing an efficient classifier model is mandatory. In this work, we present an effective hybrid multicriteria model for patients' classification. Our approach is composed of two-stage: the first one consists of generating the criteria priorities using the multi-criteria Best Worst Method (BWM) tool used to estimate a set of alternatives concerning a data set of decision criteria, while the second is based on making the patient classification using the Support Vector Machines (SVM), are controlled learning models with related learning algorithms that analyze data used for classification. This combination proposed to classify the patient's diagnostic as infected by COVID or not. In this study, we considered these criteria: fever, cough, fatigue, shortness of breath in severe cases, and the age of the patient), Predicting the right classification rapidly will reduce the number of affected people.