A system and method for region-of-interest-based bit-allocation scheme for video coding is provided. A method for encoding an image sequence of inter-frames and intra-frames includes grouping the inter-frames and the intra-frames in at least one group of pictures (GOP), and performing a frame-level bit-allocation to inter-frames and the intra-frames in the GOP. For each frame of the inter-frames and the intra-frames in the GOP, the method also includes partitioning the frame into a plurality of macroblocks, identifying macroblocks in the plurality of macroblocks as having regions of interest (ROI), and performing a macroblock level bit-allocation for the frame based on macroblocks identified as having ROI. The method further includes encoding the image sequence based on the bit-allocations, thereby producing an encoded image sequence, and outputting the encoded image sequence.