Predicting rainfall may be difficult in different places since the weather does not always follow the same pattern. However, when appropriately stored, rainwater may also be a significant supply of water. Rainfall forecasting is therefore crucial to examine this. The current study used monthly and yearly precipitation data of West Bengal (India) to anticipate rainfall. The dataset of total rainfall and relative humidity was gathered from NASA Power data-Access—Viewer from 1984 to 2022. To arrive at the most accurate rainfall prediction model, we employed five machine learning algorithms: Random Forest, Decision Tree, Naïve Bayes, Support Vector Machine, and Extra Tree Classifier. The classifiers undergo a comparative performance examination. To determine which classifier provides the most accuracy, we experimented with different splitting ratios with feature selection and 10CV. The Extra Tree Classifier has the highest accuracy, at roughly 99.16%.