Forecasting the amount and pattern of precipitation (rainfall) in a certain geographic area for a set time period, usually in the near future, is known as rainfall prediction. For many purposes, including farming, managing water resources, flood control, and disaster management, accurate rainfall projections are crucial. Establishing models of prediction that can predict future precipitation requires the use of past meteorological data and machine learning techniques. Weather patterns are often non-stationary, meaning they change over time. Machine learning models assume stationarity, which can be a limitation when dealing with climate and weather data that exhibit trends, seasonality, and long-term shifts. The proposed model uses Blended stacking model, is a technique in machine learning where multiple models are combined to improve predictive accuracy. Stacking can make rainfall prediction models more robust to changes in data distribution and weather patterns. It can adapt to varying conditions and provide consistent performance over time.Blended stacking can provide not only point predictions but also measures of uncertainty. By aggregating predictions from multiple models, it can offer insights into the variability and confidence associated with rainfall forecasts.