Energy simulation software programs help designers and energy managers to have a better understanding about building energy consumption in the design phase, but there are some occupancy related factors that are missing in these software programs. Neglecting each of these factors may lead to underestimation or overestimation of energy consumption in commercial buildings during working hours. Also, energy managers cannot have a reliable comparison about efficiency of different energy saving policies due to abandoning these factors. In this regards, the current paper proposes an agent-based approach to simulate the impact of occupancy behavior on energy consumption and emission production of commercial buildings. A model was developed and then implemented in a case study to investigate the potential energy efficiency improvement of a simulation with dynamic occupancy parameters. The results of the developed model showed 19% discrepancy with eQuest as a conventional energy simulation software program over 27 months long period.