This study investigates the prediction of water precipitation in Mexico’s metropolitan area using various machine learning techniques to enhance resilience against climate change. Data from Mexico City (CDMX), State of Mexico (EDOMEX), and Morelos, including precipitation, temperature, soil water, surface water, and underground water levels, was collected, normalized and treated using data tidying techniques. The combined datasets were analysed using linear regression, random forest, and support vector machines (SVM). Random forest models yielded the highest performance with \(\hbox {R}^{2}\) values of 0.86 for CDMX, 0.93 for EDOMEX, and 0.91 for Morelos, outperforming linear regression and SVM models. This research provides a robust methodology for predicting precipitation, crucial for water resource management and planning, contributing to climate resilience in the region. Future work may include exploring additional machine learning techniques, integrating real-time data for improved accuracy, and expanding the geographical scope to include other regions affected by climate change.