The present disclosure relates to an apparatus and method for encoding/decoding videos using prediction direction change and selective encoding. The present disclosure provides a video encoding apparatus that encodes the current block of a video, comprising a rectangle encoder for dividing and then successively encoding the current block at input into rectangular block units to output a rectangularly encoded bitstream; a square encoder for encoding the current block at input in square block units to output a squarely encoded bitstream; and an encoding selector for calculating the encoding costs of the rectangularly encoded bitstream and the squarely encoded bitstream so as to output the bitstream with a minimum encoding cost. According to the present disclosure, the prediction accuracy may be increased when encoding or decoding videos so that video encoding efficiency can be improved.