The paper introduces an innovative method for enhancing video clarity by employing a deep learning framework to address haze distortion architecture specifically designed for image dehazing: the SmokeEye. The method tackles video dehazing in a casing-by-outline way, utilizing the SmokeEye’s capacity to catch long-range conditions and complex connections inside pictures. The dehazed frames are sewed back to a video sequence, resulting in a clear and haze-free output. This approach offers several advantages like effectiveness in dehazing various video scenes, potential to outperform prior art based on image quality metrics and integration with established video processing frameworks. Our research contributes to the field of video enhancement by demonstrating the applicability of SmokeEye’s for video dehazing tasks.