This paper investigates congestion tolls under two different toll collection schemes: mainline collection and ramp co/lection. Our analyses also focus on the changes of flow pattern and system welfare for these two toll collection schemes. Congestion pricing has long been a notable research subject of traffic management and transportation policy. It is also an interesting topic for economists and transportation planners from both in theoretical and practical point of views. Most research focuses on the discussions on different pricing policies. However, few research considers the impacts of the main-line to/l and ramp to/l co/lection schemes. In this research we designed a simple but representative model to provide a quantity analysis of the changes of toll, flow pattern, and social welfare under the second best pricing policies. The cases of no toll and first best pricing policy are also discussed.