The invention relates to the technical field of camera, in particular, to a hand-held binocular fisheye 3D VR camera, which includes a housing and a lens assembly mounted on the housing. A bottom surface inside the housing is mounted with an MIC board, and a main bracket is mounted inside the housing. A rear surface of the main bracket is provided with a structural port plate, a sensor body and a sensor board bracket. A front surface of the main bracket is mounted with a cooling bracket and a main board, and one end of the rear surface of the main bracket is mounted with a speaker. The invention is provided with the MIC board and the speaker, supporting three recording modes of photographing, video recording and live broadcast.