Ensuring the preservation of vital content and maintaining the geometric integrity of an image during resizing is crucial. Seam carving is a powerful technique for achieving this, allowing for image resizing with minimal or no content loss while preserving significant elements. A “seam” in this context is an optimal connected path of pixels within an image, extending either from top to bottom or left to right, with the selection of pixels based on their energy levels. In this paper, we introduce an innovative approach based on Reinforcement Learning, specifically using Q-learning. Unlike traditional dynamic programming-based seam carving algorithms, our method employs Q-learning to make seam removal decisions through learned policies. The Q-learning agent in our approach is trained to identify the lowest energy pixels, creating an optimized seam that minimizes content loss and artifacts. The agent’s Q-values are acquired through interactions with the image environment, enabling informed decisions about pixel selection based on energy levels. Our experimental results, conducted across various image types, demonstrate the efficiency and effectiveness of our proposed method. This research makes a significant contribution to the fields of image processing and computer vision by introducing a Reinforcement Learning (Q-learning) based seam carving technique. This innovation holds promise for advancing content-aware image resizing, with applications spanning multimedia content management and beyond.