Provided are a motor control method and a device, applicable in the technical field of motor control, and particularly applicable to a controller in a motor control system. The system comprises one controller and at least two motors, and each of the motors drives one drive entity. The method comprises: acquiring a control instruction at a current time, a positional relationship of drive entities, and a system state after a previous program cycle has ended; on the basis of the acquired information, and in combination with a pre-established mapping of transition relationships and transition conditions between all system states, determining, from the control instruction at the current time, a target system state corresponding to the system state after the previous program cycle has ended, determining, according to the target system state, a target control instruction, and controlling actions of all of the motors.