Computational ghost imaging as a new imaging method, has attached the interest of many researchers. However, the practical application of computational ghost imaging is limited by the slow imaging speed of computational ghost imaging and the low quality of reconstructed images. Therefore, this paper proposes a compressed computational ghost imaging method based on array sampling, which uses the array detector to significantly increase the imaging speed of computational ghost imaging and achieve high-quality image reconstruction under low sampling times with compressed sensing. The experimental results show that the imaging method proposed in this paper achieves a speed increase of more than 4 times compared with single-pixel computing ghost imaging. Even at the sampling rate of 1.8%, high quality reconstructed images can still be obtained. Therefore, the method can promote the application of computational ghost imaging in real-time imaging, real-time detection and other fields.