A speech processing system placed in a vehicle to effectively acquire a voice of a driver includes a first microphone that is provided on a surface facing the driver of the vehicle out of a steering wheel of the vehicle, and inputs a voice of the driver, a second microphone that is provided at a position where the voice of the driver of the vehicle is blocked by at least part of the steering wheel in the vehicle, and inputs noise in the vehicle, and a noise suppressor that suppresses an estimated noise signal based on a first signal corresponding to the sound input to the first microphone and a second signal corresponding to the sound input to the second microphone, and generates and outputs a pseudo speech signal.