In the present paper, an CNN algorithm is implemented to detect whether the patients are effected with Covid, a huge pandemic that tainted a big measure of the human population in the first part of the year 2020. Moreover, the investigation introduced and employed 9 popular Convolutional Neural Networks (CNNs) for the classification of images of X-ray starting from victims with COVID, pneumonia, and healthy people. Exploration evidence showed that the CNNs can possibly identify respiratory sicknesses with high accuracy, even a large number of sample pictures should be grouped. Particularly, VGG16, Scores a 95% overall accuracy. The big values related with affectability, specificity, and exactness of Corona virus strain, shows the capacity of these models to identify positive or potentially negative covid-19 cases perfectly to reduce as much as could reasonably be expected the infection to spread into the network. As the outputs show, deciding the best model for this classification task involves major exhibition measurements. Moreover, one of the empowering results is the ability of the previously mentioned CNNs to accomplish high affectability and accuracy on the healthy class along these lines guaranteeing the minimization of false positives with respect to contamination cases which can deeply help to reduce the weight on the healthcare framework.