Covid-19 is a disease that affects the respiratory system, and causes severe symptoms that are sometimes fatal, and it is transmitted easily from one person to another, unlike other viruses. The infection with the Corona virus and the severity of infection with it are among the important issues in our lives because of the disasters it caused to humanity in all areas of life, including the economic field.Predictive models are used in different fields such as cost , risk functions, and others that depend on different forms of data. Prediction models use various forms of data, such as population demographics, past case numbers, and mobility data, to make predictions about the future spread of the virus. Some models focus on short-term predictions, while others aim to make long-term forecasts. The aim of this research was to develop a prediction model to predict the severity of Covid-19. The study utilized a sample of the staff working in the Nineveh Health Department who were infected while working. The ordinal logistic model was used to fit the best model. The optimal model was selected based on the Pearson's Chi-squared test. The sample includes 536 staff with 29 predictors. It was split into a training set and a testing set, with the 70% training set used to fit the model and the 30% testing set that is used to evaluate its performance Pseudo R-Square, the results showed that a few of the predictors were statistically significant in predicting the severity of Covid-19, therefore all the non-significant predictors were excluded, The confusion matrix of the training set shows a few misclassifications but the mean square error is high, the testing set results show few misclassifications, but the mean square error is lower The final model was able to predict severity with an accuracy of 80%. The model also identified the important predictors for the outcome e.g. extreme fatigue and runny noise