A block prediction method for decoding a picture, according to the present invention, comprises the steps of: deriving a reference block from a reference picture to predict a current block in a current picture; determining a reference illumination compensation (IC) parameter; generating an illumination-compensated reference block by applying reference IC to the reference block on the basis of the reference IC parameter; and generating a predicted sample of the current block on the basis of the illumination-compensated reference block. According to the present invention, an illumination difference between the current block of the current picture and the reference block of the reference picture can be compensated for, block prediction efficiency can be increased, and a data amount of a residual signal to be transmitted from an encoder to a decoder is reduced. Therefore, compression and coding efficiency can be increased.