Landslides are one of the most devastating natural disasters, leading to energy and destruction of buildings, loss of life, economic disruption, and damage to infrastructure. Due to the sporadic and unusual behavior of them, the early detection and accurate categorization of such events are crucial for disaster management and risk reduction. Conventional approaches for landslide identification depend on empirical models, in-situ measurements, and remote sensing methods, all of which generally have inefficiency, expenses, and limited precision. However, these approaches do not allow real-time predictions and struggle with complex terrain variations, resulting in unreliable predictions. To alleviate these issues, this study employs high-resolution satellite imagery as well as deep learning techniques. A residual learning structure based convolutional neural networks-based model that is used to identify landslide-prone areas with greater accuracy is implemented. The images are normalized and augmented applied to the data for training and testing, so that the model becomes more durable. The results show that the proposed method outperformed conventional classification techniques, with an accuracy of 97.21 percent. The results indicate that deep learning models provide a more stable automated, tunable, and scalable approach to detecting landslides which could be useful in the prevention of landslides and for mitigation.