Accidental building fires present considerable valuable information on fire phenomena and behavior. However, fire and smoke movement quantitative data cannot be obtained directly from a building fire accident. Therefore, computational methods have been widely used to reestablish the fire scenario and obtain a great deal of information for fire study. This paper utilized two types of fire simulation software, CFAST and FDS, to reconstruct a motorcycle shop fire accident in Taiwan. Important thermodynamic parameters during the fire duration were obtained through numerical simulations. The impact of natural/forced ventilation on the fire was also studied. The results showed that the more forced airflow into the building, the more time people had to evacuate to a safe place from the fire.