Based on the theory of evidence-based design, the usage behavior of air conditioners by different occupants in a case building is measured, and the typical running modes in different operating periods are put forward in this paper. By the Monte Carlo method, the annual operation sequence of air conditioners is simulated. Taking the annual average operation time of air conditioner as a checking parameter, the simulation results are compared with the actual measured results of different room types under cooling and heating conditions, separately. The error rate of the simulation results is lower than 5.5%, and thus the rationality and practicability of the simulation method are verified. The input of the stochastic operation sequence of air conditioners is helpful to work out the energy-saving and thermal environment optimization strategies based on the actual operating characteristics in building energy simulation.