An endoscopic image pickup device includes a sheath functioning as insertion section which is inserted in a body cavity, a camera head body which is coupled to a proximal end portion of the sheath in a direction perpendicular to an axis of the sheath, is rotatable about the axis of the sheath as a rotational axis, and has a center of gravity below the rotational axis, and a remote control section which is provided at a position with an extension below the camera head body. With this structure, even if an endoscope is rotated about the axis of the sheath of the endoscope, the camera head body is always kept in a constant attitude by its own weight. In addition, blurring of an image can be prevented at a time of capturing a still image by the operation of the remote control section.