Spaceborne synthetic aperture radar (SAR) image is considered to have low signaltonoise (SNR) ratio and it has difficulties in detecting airport and extracting related regions due to speckles and nontarget objects. Aiming at these problems, an algorithm scheme was proposed to implememt airport detection and extraction in spaceborne SAR image. Curvelet thresholding and Otsu’s algorithm were adopted to restrain speckle noise and to separate region of interest (ROI) from background respectively. Then an improved randomized Hough transform (RHT) based on least square method correction was introduced in airport runway detection. The whole related airport region was extracted using region growing method whose seeds were selected in runway area. The result of experiment shows that the proposed scheme is able to have the airport detected and extracted in spite of the problems of spaceborne SAR and has high precision and robustness.