In ship target detection of infrared images, the target is usually near the sea-sky line or coastline, thus the sea-sky line and coastline should be detected in advance to determine the potential area of the ship target, reduce the search range of target detection, simplify the data processing, and improve the detection speed.Aiming at the poor adaptability of the traditional sea-sky-line/coastline detection algorithm to different background images, we made an analysis to the features of sea-sky line and coastline, and proposed an algorithm for sea-sky-line/coastline detection by use of LSD line segment detection algorithm and clustering.Firstly, the local straight line contour of the image was obtained by the LSD line segment detection algorithm, and then the potential sea-sky-line/coastline area was obtained by K-means clustering.Finally, the real sea-sky-line/coastline position is determined by analyzing the texture features of the potential sea-sky line and coastline areas.The experimental results show that this method is highly adaptive to the detection of sea-sky line and coastline under various backgrounds and has high detection precision.