A learn-and-play programming method, for controlling a mechanical movement of an output shaft of a motorized toy and a domestic appliance, includes the steps of: learning the mechanical movement of the output shaft by inputting a movement data into the operation system corresponding to the mechanical movement of the output shaft and storing the movement data of the output shaft in a memorizing means; and reproducing the mechanical movement of the output shaft corresponding to the movement data in the memorizing means. Therefore, the operation system is executed through the learning program to repeatedly control the mechanical movement of the output shaft.