An electrostatic speaker including a frame; a first electrode installed at a first end of the frame; a second electrode installed at a second end of the frame to be at a predetermined distance apart from the first electrode; a suspension which is arranged between the second electrode and the first electrode and is elastically installed inside the frame; and a diaphragm assembly which is supported by the suspension and has a multilayer structure. The electrostatic speaker adopts a dual electrode structure, has the multilayer structure formed inside the diaphragm assembly, and is provided with a driving chip that applies a bias voltage to the diaphragm assembly by rectifying an audio signal after applying a driving signal that has amplified the audio signal to the dual electrodes. Thus, the invention does not require any external power, and is also capable of maintaining relatively high sensitivity by enhancing the charge density.