An encoding method of images to be divided into blocks a current picture to be coded, selecting two reference pictures from among reference pictures on a block basis, describing information which identifies each of the selected reference pictures, and performing predictive coding on the block, said method comprising: selecting, for coding a unit plural-block image formed by a plurality of blocks, one or two common reference pictures to be referenced in common among plural reference pictures; in the case where two images common reference to which reference in common, as the two selected reference pictures, describing common information which identifies the two images common reference in a common information area for the unit image are selected plural blocks, instead of describing, per block, information which identifies the two selected reference pictures; in the case where one common reference picture to be referenced in common, one of the two selected reference pictures, describing common information which identifies the one common reference picture in a common area of information unit is selected plural-block image, instead of describing, per block, information which identifies the one of the two selected reference pictures, and further describing, per block, information which identifies the other of the two selected reference pictures; generating a predictive image of a current block to be coded included in the plural image unit block, using the two selected reference pictures; and coding the current block using the predictive image.