Cloud covers are generally present in optical remote-sensing images, which limit the usage of acquired images. In previous studies, thresholding is a common and rapid method in cloud detection. However, a selected threshold is usually suitable for local study areas, and it may be failed in other cases. Besides, there are many exceptions to control, and the environment is changed dynamically. In this study, a threshold-free method based on Support Vector Machine (SVM) is proposed, which can avoid the abovementioned problems. A statistical model is adopted to detect clouds instead of a subjective thresholding-based method. According to the physical characteristics of clouds and other objects, the spectral features are appropriately designed for classification. Spatial and temporal information are also important for satellite images. Consequently, co-occurrence matrix of the Hotelling transform is used in proposed method. Experiment results demonstrate the detection accuracy of the proposed method is about 93% to 97%.