Predictive analytics if combined with machine learning approaches have the potential to play a significant role in forecasting the spread of respiratory infections. Machine learning approaches aid in the mining of data to better assess and forecast the presence or absence of covid-19 infection. In this paper, XGBoost Ensemble Classification using a fine-tuned approach is being used. Various machine learning algorithms are being analyzed and XGBoost Ensemble Classification algorithm is used to forecast and predict the covid-19 infected patient recovery chances that is the possible outcome of a covid-19 infected patient along with the prediction probability of recovery. The suggested categorization model is applied to a variety of situations. When compared to other ensemble models like random forest, hybrid ensemble and other individual classifiers such as decision tree, logistic regression, K-nearest neighbor, support vector machine and Naive Bayes, the XGBoost Ensemble Classification model gave the highest accuracy. Predicting the correct class is important. Thus, the proposed system discusses the possibility of automated prediction for covid-19 infected patient recovery chances along with the probability of recovery and the effectiveness of such predictions.