Cyclones pose a significant threat, causing widespread devastation and loss of life. Early prediction of cyclone intensity plays a crucial role in mitigating their impact. In recent years, deep learning has emerged as a promising technique for image analysis. This paper introduces a deep learning-based method for estimating cyclone strength using image datasets. By leveraging convolutional neural networks (CNNs), the proposed approach extracts essential information from satellite imagery to forecast cyclone intensity. The model is trained on a comprehensive historical dataset sourced from the National Hurricane Center’s HURDAT2 database and validated on new cyclone data. Evaluation metrics such as mean absolute error, mean squared error, and root mean squared error demonstrate the effectiveness of the CNN model in accurately estimating cyclone intensity. Training the CNN model on the historical dataset employs supervised learning, where labeled examples consisting of satellite data and corresponding cyclone intensities are utilized. Through this process, the model discerns patterns and correlations within the satellite data, enabling it to make precise predictions for unseen cyclone data.