A method performed by an apparatus of decoding a video in the unit of blocks, includes: extracting and decoding a first information corresponding to a minimum block size and a second information corresponding to a difference between the minimum block size and a maximum block size, from a bitstream; calculating the maximum block size based on the first information corresponding to the minimum block size and the second information corresponding to the difference; decoding, from the bitstream, partition information of a square current block having the derived maximum block size to identify one or more subblocks in the current block which have sizes equal to or larger than the minimum block size; decoding a prediction mode information of each of the subblocks, the prediction mode information indicating whether a subblock is inter-predicted or intra-predicted; and reconstructing each of the subblocks in the current block by predicting each of the subblocks.