A voice mixing device comprises mixing processing sections 11 to 1k provided by sampling frequency. Each of the mixing processing sections 11 to 1k adds input voice signals with the same sampling frequency to generate a first addition voice signal, converts the sampling frequencies of the first addition voice signal into sampling frequencies which can be processed by other mixing processing sections, supplies the voice signals in which the conversion of the sampling frequencies is performed to the other mixing processing sections, and adds the first addition voice signal generated in the own processing section and the first addition voice signal in which the conversion into sampling frequencies which are supplied from the other mixing processing sections and which can be processed in the own processing section is performed to generate a second addition voice signal.