This document describes modeling and predicting monthly rainfall in Tamilnadu, India using a seasonal multivariate ARIMA process. Rainfall data for Tamilnadu from 1950-2008 was analyzed using Box-Jenkins ARIMA modeling techniques. An ARIMA (1,0,1)(1,1,1)12 model was fitted to the data, which accounted for seasonality and included sea surface temperature as a predictor variable. The model provided a reasonably good fit with root mean squared error of 1.082 and mean absolute percentage error of 16.081%. The model can be used to predict monthly rainfall in Tamilnadu.