Each layer includes a data recording area and a test writing area divided into a plurality of small areas, wherein the small areas of the test writing area are recorded in advance so that other each layer can make a combination of recorded and unrecorded states with respect to the small area, where OPC is carried out, in a layer where OPC is carried out. Moreover, OPC is carried out to each small area, where a combination of recorded/unrecorded states of other each layer differs from each other, to thereby calculate, as the optimum power, an average value of the result of each OPC or a central value of the dispersion thereof.