A modular speaker includes at least an active vibrator, at least a passive vibrator, and a module housing, wherein the active vibrator and the passive vibrator are coupled to the module housing to form a vibration cavity therewithin to share with the active vibrator and the passive vibrator. When the active vibrator is operated for sound generation in response to an audio signal input, the passive vibrator is driven to vibrate through the vibration cavity for auxiliary sound generation so as to produce a point sound with full range of frequencies. The point sound with full range of frequencies produced by the modular speaker includes treble region and bass region, such that the modular speaker is able to completely restore the audio signal input as the original form to enable the listener to hear the sound quality of original raw audio signal.