A headphone sound-generating structure includes a sound-generating module and a plurality fastening elements. The sound-generating module includes a first perforated plate, a first ring-shaped spacer, a diaphragm assembly, a second ring-shaped spacer, a second perforated plate, and a second ring-shaped mounting frame, which are sequentially superposed in a receiving space defined on an ear cup. The fastening elements are extended through third mounting holes of the second ring-shaped mounting frame and fourth mounting holes of the ear cup to thereby quickly secure the sound-generating module to the ear cup to complete a headphone sound-generating structure, which can be manufactured at lowered cost and increased good yield, and allows convenient maintenance without wasting any material. A method of assembling the headphone sound-generating structure is also introduced.