The ability to make accurate weather predictions is of great importance to a number of different sectors, including agriculture, water distribution, and disaster management. In recent months, Mexico has experienced severe droughts and high temperatures, which has further highlighted the necessity for reliable weather forecasting. This article presents a weather predictor model tailored for Mexico, which employs satellite imagery and Convolutional Long Short-Term Memory (ConvLSTM) layers. The model is designed to enhance the accuracy of weather predictions by leveraging the spatiotemporal dynamics captured in satellite data. The findings indicate that the model is capable of forecasting precipitation and temperature extremes with promising results over both short and long periods. This model demonstrates significant potential for predicting climate events and is also highly relevant in addressing the current climatic challenges in Mexico, outperforming comparable models in its predictions. Moreover, it demonstrates the continued reliability of LSTMs despite the advent of more recent alternatives.