Offshore wind turbine blades are exposed to harsh environmental conditions, such as wind, rain, and salt spray, over extended periods, which can lead to structural damage. Detecting blade damage and providing effective maintenance strategies is challenging, and these issues can result in power losses and significant economic costs. To improve detection efficiency and accuracy, this study employs various object detection algorithms to identify and localize offshore wind turbine blade faults. The experimental dataset includes three types of wind turbine blade damage: crack, skin debonding, and lightning strike. By comparing the performance of different object detection models based on key metrics such as detection accuracy, mAP (mean Average Precision), and model parameters, and evaluating detection performance using actual offshore wind turbine blade damage data, the experimental results show that YOLO11 performs the best across all aspects. It excels in both detection accuracy and inference speed, demonstrating significant practical value in offshore wind turbine blade damage detection.