The invention discloses a voice system which comprises a memory, a digital signal processor, a coding and decoding device, a loudspeaker, and a detection microphone. The memory is used for storing sound source data. The digital signal processor is connected with the memory and used for processing the sound source data into play data. The coding and decoding device is connected with the digital signal processor and used for decompressing the play data into play signals. The loudspeaker is connected with the coding and decoding device and used for playing signals into sound source sounds. The detection microphone is connected with the coding and decoding device and used for recording the sound source sounds as detection signals. The coding and decoding device is used for compressing the detection signals into detection data. The voice system also comprises a detection device which is connected with the memory, the digital signal processor, the coding and decoding device and the loudspeaker. The detection device is used for comparing the detection data with the sound source data and the playing data when the sound source sounds have noise, and comparing the detection signals with the play signals to detect whether the sound source data, the digital signal processor, the coding and decoding device and the loudspeaker have faults or not. The invention also provides an abnormal sound detection method and an electronic device.