A decoding buffer capacity necessary for decoding N pieces of digital data, including a picture, audio, a character and another form of data, is found beforehand. Then the found capacity is compared with a predetermined value. M pieces of digital data are selected out of the N pieces of digital data so that a total buffer capacity necessary for decoding the M pieces of digital data is not more than the predetermined value. Then the M pieces of digital data are multiplexed and supplied as one multiplexed data. This data multiplexing method enables systems having different buffer capacities in their own decoders to exactly decode digital data by changing decoding buffer capacities that are predetermined. As a result, a transmission free from missing an important data is realized.