The proposal presents an Intelligent Image Compression Broker Based on Objective with Feedback. A novel structure is applied to receive the input image/video, its metadata (environment and camera parameters), and the feedback from the compressed image quality evaluation to optimize the tradeoff between the compressed image quality and its size (in terms of bits the compressed image occupied). We also propose to apply machine leaning techniques to train the compression method and parameter selection broker, which also accepts real data as a part of the reference as the image quality and the application requirement match each other. Due to the application of the novel system structure, the image compression broker has the capability to continuously adjust the compression algorithm and parameters, and thus the image compression performance is optimized.