Disclosed are a method for controlling a sound box, a sound box and a non-transitory computer readable storage medium. The method is applied to the sound box in a sound box system. The sound box system includes at least two sound boxes, each sound box includes at least two first sensors, the at least two first sensors are provided on opposite sides of the sound box, each first sensor is configured to transmit and receive preset signals, and the method includes following operations: obtaining, by the sound box, a connection status with other sound boxes; when the sound box is wirelessly connected to the other sound boxes, obtaining first sensing data detected by a first sensor; determining target channel information according to the first sensing data; and outputting audio according to the target channel information.