In the present invention, at least two component signals that represent luminance and a color difference in an original signal having a predetermined image format are encoded/decoded. This encoding device (10) is equipped with: a local decoding section (17-1, 17-2, 18-1, 18-2, 23-1, 23-2) that generates a locally decoded signal of a second component signal; a luminance-color difference prediction unit (4) that generates a parameter for normalizing the second component signal by comparing the original signal of a first component signal not yet encoded with the locally decoded signal of the second component signal in order to generate a luminance-color difference prediction signal; a prediction signal generation unit (22-1) that generates a prediction signal for an intra-screen prediction or a prediction signal for an inter-screen prediction regarding the first component signal; a synthesis unit (2) that performs a weighted addition of the prediction signal and the luminance-color difference prediction signal in order to generate a corrected prediction signal for the first component signal; and an encoding section (13-1, 14-1, 15-1, 16-1) that generates and encodes a difference signal between the corrected prediction signal and the original signal of the first component signal.