This paper addresses a scheduling method of air conditioner operation using workers daily action plan. In order to run an air conditioner for both energy saving and each worker's comfort, the proposed method decides how high and when to set the temperature as operations. Comfortable temperature for workers is decided by using PMV(Predicted Mean Vote), and the temperature and power consumption are estimated by indoor environment simulator. To make an operation schedule considering both each worker's comfort and power consumption is realized by reinforcement learning. Experimental results show that our proposed method can generate a schedule to satisfy each worker's comfort at all times and reduce power consumption.