Abstract: With the development of high resolution imaging techniques, the volume of data adds serious challenges to the compression, storage and transmission of images. Recently, compressed sensing theory adds new solutions to image compression since the ability of reconstructing original signals with a small amount of observation values.The sparsity of the signal is the premise of the application of the compressed sensing theory, so the sparse representation of the data is the key step in the compression of image. The key of sparse representation is the dictionary, and the main dictionary models are synthesis model and analysis model. Along with the extension of the application of the dictionary learned through the synthesis model in the image compression, the time-consuming of the image in the sparse representation becomes a key factor restricting the efficiency of the system. Therefore, in view of the defect of the synthesis model in the application, combined with the advantages of the analysis model in the process of sparse representation, we proposed an image block compression model based on analysis dictionary(ALDBCS). In this model, firstly, a dictionary is learned through the prior data set, and then in order to reduce the cost of the sparse representation, the dictionary is introduced to the process of image compression. The standard testing library of natural images is used as testing images, time-consuming and reconstruction quality are taken as evaluation criterions, the experimental results proves that the ALDBCS model can not only improve the quality of image reconstruction, but also reduce the time consuming of image compression.