The invention relates to a method for decoding a stream of coded data representative of a video comprising at least one first view and a second view different from the first view, in which a reference image of the second view is previously reconstructed. For at least one block of a current image of the first view, termed the current block, a motion information item associated with said current block and representative of a displacement of said current block in said previously reconstructed reference image, is obtained (E64). A prediction block is obtained (E66) on the basis of said previously reconstructed reference image of said second view and of said motion information item obtained. The current block is reconstructed (E68) on the basis of the prediction block, of an decoded prediction residual information item (E650) on the basis of the stream of coded data and of a difference information item calculated between a first block associated with the first view and a second block associated with the second view and decoded (E67) on the basis of the stream of coded data.