Disclosed is an apparatus for decoding motion information in merge mode. The apparatus for decoding motion information in merge mode discloses a merge mode motion vector decoding unit configured to generate motion information using available spatial and temporal merge candidates when a motion information encoding mode of a current block indicates a merge mode; a prediction bock generating unit configured to generate a prediction block of the current block using motion information; and a residual block generating unit configured to perform an entropy-decoding process and an inverse-scanning process on residual signals to generate a quantized block, and to perform an inverse-transforming process on the quantized block to generate a residual block. Furthermore, a reference picture index of the temporal merge candidate is set to 0, and a motion vector of the temporal merge candidate is selected among a first merge candidate block and a second merge candidate block based on a position of the current block within a slice or a largest coding unit, and the motion vector of the second merge candidate block is selected as the motion vector of the temporal merge candidate if the current block is adjacent to a lower boundary of the largest coding unit.