Object detection is a necessary part of the intelligent driving system, especially in some severe weather conditions like haze. However, in haze weather, images acquired by the camera are degraded, which severely impacts the subsequent detection process. The radar performance is almost unaffected by haze. Based on the above background, we focus on the multi-sensor fusion dehazing algorithm applied to vehicle detection in hazy road environments, using radar data to provide information supplements for dehazing. First, we propose a radar-camera fusion dehazing algorithm based on the atmospheric scattering model. The radar detection information is used to provide direct and accurate transmission estimation and haze removal region of interest for the dehazing algorithm, so as to achieve precise dehazing effects focusing on dynamic vehicle targets on the road. Second, we apply YOLOv5 to haze-removal images. The proposed fusion dehazing algorithm is tested on the synthetic haze road environment and the real-world dataset, respectively. The average precision of the object detection task is used as the evaluation metric. Finally, experimental results demonstrate that the proposed algorithm performs favorably against the state-of-the-art dehazing algorithms.