The present disclosure relates to a video encoding/decoding apparatus and method using a multi-dimensional integer transform. The video encoding method of the present disclosure includes: decomposing subblocks in units of a rectangular block and successively encoding rectangular block units to output a rectangularly encoded bitstream of each subblock; encoding the subblocks in units of a square block to output a squarely encoded bitstream of each subblock; calculating an encoding cost involving at least one rectangularly encoded bitstream; calculating an encoding cost not involving the rectangularly encoded bitstream; and on the basis of the calculations, using one or more of the squarely encoded bitstream and the rectangularly encoded bitstream to generate and output the bitstream of the macroblock. According to the disclosure, since the coding scheme for encoding or decoding a video with a higher efficiency is selectively determined, the prediction accuracy can increase and thereby improve the video encoding efficiency.