A modular loudspeaker includes a plurality of chambers and a plurality of vibration plates. Each vibration plate is fixed to two chambers, to cause the chambers to be secured to each other in a certain orientation by the vibration plates. A vibrator is attached to each vibration plate. A processor and a wireless communication unit are received in each chamber. The vibrator attached to each vibration plate is electrically connected to the processor, and the processor is configured to control the vibrator to vibrate when the loudspeaker receives a wireless audio signal via the wireless communication unit, thereby causing the vibration plate to vibrate and output sound.