It is known that in every quotation of exchange rates, there are two currencies involved. Exchange rate is defined as the price of a unit of foreign currency in terms of a domestic currency. This research aims to provide the best Autoregressive Integrated Moving Average (ARIMA) model that forecasts the Philippine Peso - United States Dollar (PHP-USD) exchange rates using historical average monthly rates from January 2009 to December 2016. Forecasted values from different ARIMA models are compared in terms of errors with actual data values. The best ARIMA model produced is ARIMA(0, 2, 2), which yielded the lowest values for the different performance measures used.