In residential building, energy consumption of air conditioning is proved to be largely influenced by the behaviors of occupants. Therefore, the key to predicting the energy consumption of Room Air Conditioner (RAC) is to accurately understand how occupants control RAC. The data on the usage of RAC was obtained mostly by investigations in previous studies, resulting in a small sample size and short duration. Although the American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE) recommends a uniform occupancy schedule, occupants’ behavior patterns vary by region and time. Based on the monitoring data in summer in Chongqing, the significant factors affecting the energy consumption of the RAC are statistically analyzed, which include the setting temperature and the daily usage rate. Further, the behavioral pattern of air conditioning usage is analyzed to investigate the impact of daily mean outdoor temperature in the paper, which meets the "L-curve". The result shows that occupants have a low tolerance to the indoor thermal environment and are likely to use air conditioner even when the outdoor temperature is at a lower level. Then the optimized energy consumption simulation model based on the traditional simulation model and the energy consumption prediction model are proposed to compare with the measured energy consumption. It is proved that the results of the optimized energy consumption simulation model adopted the daily usage rate and the energy consumption prediction model based on the setting temperature, outdoor daily mean temperature, the length of daily operating time turn out to be closer to the measured value than the traditional energy consumption simulation model.