From April of 2023 to March of 2024, this study will forecast the quarterly performance of international banks in India. The study utilizes quarterly data from March 2019 through March 2023. The Autoregressive Integrated Moving Average (ARIMA) model, a common machine learning approach for time series forecasting, is used to accomplish this goal. At first, the dataset is investigated and preprocessed to make sure it can be used for modeling. Next, we apply the ARIMA model to the historical data in order to project outcomes for the subsequent four quarterly time intervals. Due to its ability to minimize mistakes and maximize prediction accuracy, the ARIMA model's optimal parameters are found to be (0, 1, 0) after extensive testing. The Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) are computed to verify the ARIMA model of choice. Useful information about the model's goodness-of-fit and whether or not it reflects the underlying patterns and dynamics of the data may be gleaned from these metrics. The research checks the Autocorrelation Function (ACF) and Partial Autocorrelation Function (PACF) to make sure the time series is stationary. The ARIMA model may be improved with the use of these diagnostic tools by locating any residual non-stationary components in the data. Researchers hope that these findings would help stakeholders and decision-makers in India better understand the likely future performance of international banks in the country. In addition, this demonstration of the usefulness of machine learning techniques for forecasting financial time series data is exemplified by the use of ARIMA modeling techniques in this scenario.