Provided is an optical recording method for recoding information on a phase-change optical recording medium at a low linear velocity by use of a divided pulse method. When forming a plurality of recording marks by use of a constant divided pulse number m, a light irradiation time period of recording light for recording a recording mark having a time length of nT is divided in the order of {.alpha..sub.1T, .beta..sub.1T, .alpha..sub.2T, .beta..sub.2T, . . . , .alpha..sub.iT, .beta..sub.iT, . . . , .alpha..sub.mT, .beta..sub.mT} (in which m is a natural number representing a divided pulse number, .alpha..sub.i (1.ltoreq.i.ltoreq.m) is a real number greater than 0, .beta..sub.i (1.ltoreq.i.ltoreq.m-1) is a real number greater than 0, and .beta..sub.m is a real number equal to or above 0). Then, the recording light having increased recording power Pw.sub.i for recording the shortest recording mark is applied so as to provide m satisfying a relation that an average value Pw.sub.short of the recording power Pw.sub.i at the time periods of .alpha..sub.iT for forming the shortest recording mark is greater than an average value Pw.sub.long of the recording power Pw.sub.i at the time periods of .alpha..sub.iT for forming the longest recording mark (Pw.sub.short>Pw.sub.long).