A Severe Acute Respiratory Syndrome (SARS) spread model based on System Dynamics (SD)approach is implemented to do the quantitative analysis of the critical importance of control policies such as early report, early quarantine, and early hospitalization to control further SARS epidemic spread. Using this model, a case study on Beijing SARS epidemic situation is conducted. The simulation results show that the onset-to -hospitalization interval and effective quarantine are two key factors in SARS transmission. Furthermore, the shortening of onset-to-hospitalization interval should be given the highest priority. In addition, a period based simulation is carried out to analyze SARS spread and control effectiveness in Beijing. The simulation outcome demonstrates that various control measures after April 20,2003 in Beijing and the public holidays starting from May lst to May 5th have a significant impact on containing the epidemic situation. The most dangerous days, around April 27, when the number of uncontrolled SARS patients in the society reached at its peak, can also be calculated. The SARS infectiousness number newly reported in Beijing has been at a low level since May 20. But simulation experiments describe that it may fluctuate with import of SARS infectiousness from outside Beijing, weakening of quarantine, more social activities, and so on.