Introduction: The current outbreak of COVID-19 has caused the world to stop and go under lockdown and has quickly grown to become a pandemic. The clinicians and scientists in medical industries are observing the pandemic for screening the COVID -19 virus in a person. Objective: In these trying times, we thought of analysing the trends in COVID-19 cases in the USA, India and Brazil using Several Time Series, Machine Learning and Ensemble Learning algorithms to check out the trends. Methods: In this paper, Auto-Regressive Integrated Moving Average (ARIMA) and Long Short-Term Memory (LSTM) under Time Series, Support Vector Regression (SVR) and Linear Regression under Machine Learning algorithms and Random Forest Regression, XGBoost and AdaBoost under Ensemble Learning were discussed. Results: After analyzing the results of all the algorithms, we observed that ARIMA and LSTM were performing better than the others for Time-Series Forecasting. This study would be valuable for medical Researchers and the Government in the future. Conclusion: Seven models, namely, ARIMA and LSTM models under time-series analysis models, support vector regression and linear regression under machine learning models and random forest regression, XGBoost and AdaBoost under ensemble learning were discussed. We first looked at the sample fits and then successfully forecasted the trends for the new cases, deaths and total cases for the next 30 days in the two countries with the highest number of cases, namely, India and the US. From the resultant graphs and table values, we could infer that overall, time-series models like ARIMA and LSTM perform the best in situations like these where data is continuous and forms a series