Providing quality-of-service (QoS) for video delivery in heterogeneous wireless networks is a challenging problem. To have QoS provision at application level, effective congestion control is very necessary. We investigate this problem and present a network-aware scalable video congestion control mechanism in heterogeneous wireless networks. The mechanism not only adapts to the dynamical network condition, but also improves the perceived visual QoS. In this control mechanism, first, each layer of the scalable video is given different protection in terms of packet loss and delay. Secondly, we monitor or model the bottle link to estimate the available network bandwidth dynamically. Thirdly, the QoS mapping module translates the QoS exponents to the effective capacity according to the effective capacity theory. Lastly, the congestion control mechanism adjusts the scalable video according to the compare result between the estimated bandwidth and the requested effective capacity. Finally, simulation results demonstrate effectiveness of our congestion control mechanism.